From patchwork Mon May 11 12:26:54 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lukas Straub X-Patchwork-Id: 11540595 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 6285492A for ; Mon, 11 May 2020 12:31:21 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 342BE20746 for ; Mon, 11 May 2020 12:31:21 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=web.de header.i=@web.de header.b="RToi7df7" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 342BE20746 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=web.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:45380 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jY7aa-0008Lo-6Y for patchwork-qemu-devel@patchwork.kernel.org; Mon, 11 May 2020 08:31:20 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:37672) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7Wo-0006BP-1M for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:26 -0400 Received: from mout.web.de ([212.227.17.11]:35071) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1jY7Wn-0005xt-6W for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:25 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=web.de; s=dbaedf251592; t=1589200017; bh=D/TPeLNov0t6rmkFP3E5KzxL/A8EmZiUQlUmOgNwPxk=; h=X-UI-Sender-Class:Date:From:To:Cc:Subject:In-Reply-To:References; b=RToi7df7ENuhfEuHE+5ZtG+4Aa3g1p/xmJ730tewZIYz6lrkBYPkIjvCnZjbOht40 eDWFuVPH2B+DaM1n4xUs4ae50zT4kdLN3c6LntxSEC3NsPrfjwO2c+Wie5uYlTKsTk eDGfPPq2olTQ3aqq7b9UB8av09po/OTZ7H0tPEPo= X-UI-Sender-Class: c548c8c5-30a9-4db5-a2e7-cb6cb037b8f9 Received: from luklap ([89.247.255.192]) by smtp.web.de (mrweb101 [213.165.67.124]) with ESMTPSA (Nemesis) id 0Lj2QM-1iyW3a3W8s-00dFOX; Mon, 11 May 2020 14:26:56 +0200 Date: Mon, 11 May 2020 14:26:54 +0200 From: Lukas Straub To: qemu-devel Subject: [PATCH 1/5] block/quorum.c: stable children names Message-ID: <9df6b3723ec30cb749ceaa555d82a29a6d79496d.1589199922.git.lukasstraub2@web.de> In-Reply-To: References: MIME-Version: 1.0 X-Provags-ID: V03:K1:TIVVxpWm1nKnNCw1LaVnWP+BopzU4pmXBGarktqhyf4VbL2PC1h fX0lSGW74sPyoioVMcuboJRiUuwmLGtga/dGV+fvAvSAlqTFvuamSSCsMI/rECblSUgda0m HWIHAkEoFD9QxzTyMmLa3jmLRAY+NB06XIjLJ3DA5XGQPTpqkfhHyoacpmWQH4XRvyqDCTP QlAGwS5V7paOlx7lvDDWA== X-UI-Out-Filterresults: notjunk:1;V03:K0:jMQYuGskO74=:QsZjYQRoZhUU8ABPzPUtec omlYvTjl8TPNKEwLXvN6EhEaeKoQdPsAUwqzYgnWB1/4mUMJYG5CGhThDNLNyEpGFqErr9cEZ IDM9vqG4sWFmXyCltSRycVEppt7eK69k/ZfkUIvXdbn8Jq8FFitRW5NGLrW9qkXkkRdwXsos9 Uaq9ygBp9W5bpju5vmHMSbUHaB1m/FZXOZyS5JdesYw2PUcm7i/5gCDw4LsLoi9oIYEtHqtvD kD92rmqJxNn/aDGWzDw6MqdRfv8YnYAC5jmI/zPH0Qj7tVYpw2HcJBjj1oOyF3qA0cF91DAPY U9AJhfDQmzereH3OaO0MDyZfLp0mswV4e8GqOjNsJEK4Q9t+8OD+vKCfXWxYBQLI1au17PATz cNJH1Ln2X65a+YHyYZsts2xMODCTUeOXuoAqul/gy2I8BrQ0zjZV4ulg/JHVQBGE8AzqAeULQ NNky8E7MJP0wKpMoWfnzMuixjkAOpkONgWu4NtpZznXGbmDTJHzsZxyrENqUXJ88sUXhsynTY V2WGJ2UnKY+j6huoTq8fn38VvEUMk+RyON2tHY8plcJTUQeCPfP9zu250DDD8GdgLsfJxZ9UD /9Kaw4y20uJGI/1xba0zhLOtdwAVoMBvGGxuKYQYcNsJy6Ws0MMdbGFpYkFpO+baX24ZOBDFs Rx3QciJ7DfaJeVw8rXcPgq94tDjqxFRASIbJM2Fe8xyS0O2pibgTcbGRLcLIPdVBC5eCZBkoj kwhGqvsNutYoT/c6SSlTcI8o4T2T/jHrd1wuTcIi142A3h3D8OSkljabcyrCbElD4DdyrQ+qP vH7acm3HNnwExF+tfsQnzTG/rkX2YG8ntB5Tq78o/W9FAWVVVPVv2m2Ro/jWuzhO1DgsXmFqN tZRnpXpqRK9CGDA24B7wTv1aZB3h/RkYZ2HKzPXkelio5TXrR67wk6PRlS+w39X/3qGbFDFP7 295SHdXqgYOA2QNi83GCMNEiI1V9snRtUPc1WL2+BxhsHlvHiqBY4KhrcmgvcPmk8Cm/sB3or UTUuiiGdcL4+63eds020bsN649lA4QLj6BI0k3uOYf8iPH/nMAaVIzPYoqJNx1HQ/k30bxjVn YhdjuTAZDmtx18uLy7bBTkwak7zH78JjTGU9pWUrcHbdpgp/2c3xYfRKYvTZxPbOfp+m/NwOz Ejkf9LxQGQx0MdBLGzbvtsXBXFj3QFGAMtln2yUmihCalH5zDhbnOtccevbGublD4rAdL3/YN PE2J8UK+PBO3sW//s Received-SPF: pass client-ip=212.227.17.11; envelope-from=lukasstraub2@web.de; helo=mout.web.de X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/11 07:14:43 X-ACL-Warn: Detected OS = Linux 2.2.x-3.x [generic] X-Spam_score_int: -24 X-Spam_score: -2.5 X-Spam_bar: -- X-Spam_report: (-2.5 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_ENVFROM_END_DIGIT=0.25, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_PASS=-0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Zhang Chen , Alberto Garcia , "Dr. David Alan Gilbert" Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" If we remove the child with the highest index from the quorum, decrement s->next_child_index. This way we get stable children names as long as we only remove the last child. Signed-off-by: Lukas Straub Reviewed-by: Zhang Chen Reviewed-by: Alberto Garcia --- block/quorum.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/block/quorum.c b/block/quorum.c index 6d7a56bd93..acfa09c2cc 100644 --- a/block/quorum.c +++ b/block/quorum.c @@ -29,6 +29,8 @@ #define HASH_LENGTH 32 +#define INDEXSTR_LEN 32 + #define QUORUM_OPT_VOTE_THRESHOLD "vote-threshold" #define QUORUM_OPT_BLKVERIFY "blkverify" #define QUORUM_OPT_REWRITE "rewrite-corrupted" @@ -972,9 +974,9 @@ static int quorum_open(BlockDriverState *bs, QDict *options, int flags, opened = g_new0(bool, s->num_children); for (i = 0; i < s->num_children; i++) { - char indexstr[32]; - ret = snprintf(indexstr, 32, "children.%d", i); - assert(ret < 32); + char indexstr[INDEXSTR_LEN]; + ret = snprintf(indexstr, INDEXSTR_LEN, "children.%d", i); + assert(ret < INDEXSTR_LEN); s->children[i] = bdrv_open_child(NULL, options, indexstr, bs, &child_format, false, &local_err); @@ -1026,7 +1028,7 @@ static void quorum_add_child(BlockDriverState *bs, BlockDriverState *child_bs, { BDRVQuorumState *s = bs->opaque; BdrvChild *child; - char indexstr[32]; + char indexstr[INDEXSTR_LEN]; int ret; if (s->is_blkverify) { @@ -1041,8 +1043,8 @@ static void quorum_add_child(BlockDriverState *bs, BlockDriverState *child_bs, return; } - ret = snprintf(indexstr, 32, "children.%u", s->next_child_index); - if (ret < 0 || ret >= 32) { + ret = snprintf(indexstr, INDEXSTR_LEN, "children.%u", s->next_child_index); + if (ret < 0 || ret >= INDEXSTR_LEN) { error_setg(errp, "cannot generate child name"); return; } @@ -1069,6 +1071,7 @@ static void quorum_del_child(BlockDriverState *bs, BdrvChild *child, Error **errp) { BDRVQuorumState *s = bs->opaque; + char indexstr[INDEXSTR_LEN]; int i; for (i = 0; i < s->num_children; i++) { @@ -1090,6 +1093,11 @@ static void quorum_del_child(BlockDriverState *bs, BdrvChild *child, /* We know now that num_children > threshold, so blkverify must be false */ assert(!s->is_blkverify); + snprintf(indexstr, INDEXSTR_LEN, "children.%u", s->next_child_index - 1); + if (!strncmp(child->name, indexstr, INDEXSTR_LEN)) { + s->next_child_index--; + } + bdrv_drained_begin(bs); /* We can safely remove this child now */ From patchwork Mon May 11 12:26:59 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lukas Straub X-Patchwork-Id: 11540611 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id C4B74912 for ; Mon, 11 May 2020 12:37:16 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 3A40D206D6 for ; Mon, 11 May 2020 12:37:16 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=web.de header.i=@web.de header.b="GUXFlATa" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 3A40D206D6 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=web.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:52098 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jY7gJ-00043i-2y for patchwork-qemu-devel@patchwork.kernel.org; Mon, 11 May 2020 08:37:15 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:37682) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7Wr-0006Cm-IA for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:29 -0400 Received: from mout.web.de ([212.227.15.4]:55515) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1jY7Wo-0005yR-2B for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:29 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=web.de; s=dbaedf251592; t=1589200021; bh=lTo/JhFEDXuaNPIN05x18j+fNnceZj6xEwdn8WEqPAw=; h=X-UI-Sender-Class:Date:From:To:Cc:Subject:In-Reply-To:References; b=GUXFlATa8ommnNtWM7CGrbCBwipX8X5m2kyb6/eElw7Y0TGA5P/je++dlYzr+OAd0 ptIANqKfruguxBsfBSKG8KC8DPZLD4+6N0v39NrkPnx6klJ9oH8ITF9mkdmn8h7LdT xZOyPFchplpODPyX/WUlCLGK4DIOdzlRCIzXptt0= X-UI-Sender-Class: c548c8c5-30a9-4db5-a2e7-cb6cb037b8f9 Received: from luklap ([89.247.255.192]) by smtp.web.de (mrweb003 [213.165.67.108]) with ESMTPSA (Nemesis) id 0MCZfe-1jPuKM2IWg-009SMB; Mon, 11 May 2020 14:27:00 +0200 Date: Mon, 11 May 2020 14:26:59 +0200 From: Lukas Straub To: qemu-devel Subject: [PATCH 2/5] colo: Introduce resource agent Message-ID: In-Reply-To: References: MIME-Version: 1.0 X-Provags-ID: V03:K1:YPEVo7+1CZztij583vqrkWMCJHFMxR/TCg8LguO2E7ugsHVKDIp vR1PbhLr0gzCWzln2sEB4Hs5dqzhT01XphsC8xcRkWuxdi2AAYMPgNXwYVdI7VMCBLW0xix lXXgNbOUeJM6/pgwyBrVdcBN+UU4psdfcX6pX0G/0jtkO8nK0IKhetFAoaNvRFr/1iO2egN jZ1jrgI94Rw2QF1N199Tg== X-UI-Out-Filterresults: notjunk:1;V03:K0:LXUoRgrMmNo=:5Xyc2ZMIydu9WPWEIFpXYl Nf/OIdy35fqHyIjp0BovuDn67NTyUPN1wrYWy/u7A8U6ECroPFSz6vck1jN/06Jd4eKj6r+yP hsVlLR9ENF0eABUfIscQkpRta9P+rBQtITAWwBvJdbslYnBsNg6RkFcBMUFUs8fJfqdXykjrH /UUx9vouRwHF5WmPm0TSlM1uhK4ivAHSRYUf9HLcvEV2UFg0TOjO7MWyF6Ml6KDfmwlZOE8TL LnDi45aTuXEcSNjQdRE9jwjOwv51TKkgUHZBf6pAxfF0IZ3s8+5JBbqWQwPoXNEoBLe0avfEm PmsSuCLtJP5dCeWGfmTbvEXkitaZYEfLWd9YifE0TJY2XoiOEaet5JGMO986sPRn1h87OOmuT rIM+W70/3VIKXPPwj320Hweh8HJoMhPP3bGLhp/ZnP4Rv5ZC6yAFfMFjeeHXIs0vZlJnGfkBD VuI4kFRALeKVXYyuJbgQLCahrYc5eZZor0pHWpn1/m77UqhUAmBHxHrDfd8pWVkJiLbl3u2N1 gunpAXZXSGm+e6e8T9D/2+q+75V9/4bjF19Au/furX/36Uke5/nXwPkqvcn5mnz58Icn1yBvk dtj4Wk5vWFEIahG6s929ffIN/yhEKWkRFY0eIf9AmWYMixIKpHcFvIzSj5ofdu0JRn/7Me14g 2PernnHt0yloa9PyIv2QCNXJsV+zepv9Zb+tTTImt726tPARfei+qxp4TQL5dgEEKsQlKZyrW eVrrSSVbtuX23iE+BONC5zaH9bHRtMUHWSspPNyRrGehxWEByLu4NJGL6T8+cA6vsMC2NpyNN tL55JoDDnZQhosx7GMIFQpZJGUKkiW+dsBJEa9gh2OuTMvxzjWpFUyA/8n2msr2Sb7uiBKZ15 P13qSe20SsWZH4aqz/+WL8SFJV9kgSZIM+L5ti84L8OIyRuEBkRh3StCnnb57kvqHrwVPHV+b mvghXmpXdJN92fVOOGhvfadx9Ra1GnykvxBloTEZwtt4a6qsnQuZCRNOKQ2mT0S3qdKzpM7vq N0RrwS59X5RsKs3mLpmk/SxmJ/J1+DlcbVFSw0YlBODAhNmIVtotiRAR8zuA5Ysil2Tv8XPb7 igRLbDrqN1iy3a0+HiU2jYJxhrbHFP+pkjM5iFjR5+exhcYkFPmy1khR5Rrd1r8r0lxJIhOro gp6MHE0nnBFdDeZfZir1QrVJd2zF2nbytcZ2l2tLFwaEpMTT4LodEG951G8/lbrkLf8ES5i2r xWxcTBM2j0dP466dl Received-SPF: pass client-ip=212.227.15.4; envelope-from=lukasstraub2@web.de; helo=mout.web.de X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/11 07:14:42 X-ACL-Warn: Detected OS = Linux 2.2.x-3.x [generic] X-Spam_score_int: -24 X-Spam_score: -2.5 X-Spam_bar: -- X-Spam_report: (-2.5 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_ENVFROM_END_DIGIT=0.25, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001, WEIRD_QUOTING=0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Zhang Chen , Alberto Garcia , "Dr. David Alan Gilbert" Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" Introduce a resource agent which can be used to manage qemu COLO in a pacemaker cluster. Signed-off-by: Lukas Straub --- scripts/colo-resource-agent/colo | 1429 ++++++++++++++++++++++++++++++ 1 file changed, 1429 insertions(+) create mode 100755 scripts/colo-resource-agent/colo diff --git a/scripts/colo-resource-agent/colo b/scripts/colo-resource-agent/colo new file mode 100755 index 0000000000..fbc5dc2c13 --- /dev/null +++ b/scripts/colo-resource-agent/colo @@ -0,0 +1,1429 @@ +#!/usr/bin/env python3 + +# Resource agent for qemu COLO for use with Pacemaker CRM +# +# Copyright (c) Lukas Straub +# +# This work is licensed under the terms of the GNU GPL, version 2 or +# later. See the COPYING file in the top-level directory. + +from __future__ import print_function +import subprocess +import sys +import os +import os.path +import signal +import socket +import select +import json +import re +import time +import logging +import logging.handlers + +# Constants +OCF_SUCCESS = 0 +OCF_ERR_GENERIC = 1 +OCF_ERR_ARGS = 2 +OCF_ERR_UNIMPLEMENTED = 3 +OCF_ERR_PERM = 4 +OCF_ERR_INSTALLED = 5 +OCF_ERR_CONFIGURED = 6 +OCF_NOT_RUNNING = 7 +OCF_RUNNING_MASTER = 8 +OCF_FAILED_MASTER = 9 + +# Get environment variables +OCF_RESKEY_CRM_meta_notify_type \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_type") +OCF_RESKEY_CRM_meta_notify_operation \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_operation") +OCF_RESKEY_CRM_meta_notify_key_operation \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_key_operation") +OCF_RESKEY_CRM_meta_notify_start_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_start_uname", "") +OCF_RESKEY_CRM_meta_notify_stop_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_stop_uname", "") +OCF_RESKEY_CRM_meta_notify_active_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_active_uname", "") +OCF_RESKEY_CRM_meta_notify_promote_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_promote_uname", "") +OCF_RESKEY_CRM_meta_notify_demote_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_demote_uname", "") +OCF_RESKEY_CRM_meta_notify_master_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_master_uname", "") +OCF_RESKEY_CRM_meta_notify_slave_uname \ + = os.getenv("OCF_RESKEY_CRM_meta_notify_slave_uname", "") + +HA_RSCTMP = os.getenv("HA_RSCTMP", "/run/resource-agents") +HA_LOGFACILITY = os.getenv("HA_LOGFACILITY") +HA_LOGFILE = os.getenv("HA_LOGFILE") +HA_DEBUG = os.getenv("HA_debug", "0") +HA_DEBUGLOG = os.getenv("HA_DEBUGLOG") +OCF_RESOURCE_INSTANCE = os.getenv("OCF_RESOURCE_INSTANCE", "default-instance") +OCF_RESKEY_CRM_meta_timeout \ + = os.getenv("OCF_RESKEY_CRM_meta_timeout", "60000") +OCF_RESKEY_CRM_meta_interval \ + = int(os.getenv("OCF_RESKEY_CRM_meta_interval", "1")) +OCF_RESKEY_CRM_meta_clone_max \ + = int(os.getenv("OCF_RESKEY_CRM_meta_clone_max", "1")) +OCF_RESKEY_CRM_meta_clone_node_max \ + = int(os.getenv("OCF_RESKEY_CRM_meta_clone_node_max", "1")) +OCF_RESKEY_CRM_meta_master_max \ + = int(os.getenv("OCF_RESKEY_CRM_meta_master_max", "1")) +OCF_RESKEY_CRM_meta_master_node_max \ + = int(os.getenv("OCF_RESKEY_CRM_meta_master_node_max", "1")) +OCF_RESKEY_CRM_meta_notify \ + = os.getenv("OCF_RESKEY_CRM_meta_notify") +OCF_RESKEY_CRM_meta_globally_unique \ + = os.getenv("OCF_RESKEY_CRM_meta_globally_unique") + +HOSTNAME = os.getenv("OCF_RESKEY_CRM_meta_on_node", socket.gethostname()) + +OCF_ACTION = os.getenv("__OCF_ACTION") +if not OCF_ACTION and len(sys.argv) == 2: + OCF_ACTION = sys.argv[1] + +# Resource parameters +OCF_RESKEY_qemu_binary_default = "qemu-system-x86_64" +OCF_RESKEY_qemu_img_binary_default = "qemu-img" +OCF_RESKEY_log_dir_default = HA_RSCTMP +OCF_RESKEY_options_default = "" +OCF_RESKEY_disk_size_default = "" +OCF_RESKEY_active_hidden_dir_default = "" +OCF_RESKEY_listen_address_default = "0.0.0.0" +OCF_RESKEY_base_port_default = "9000" +OCF_RESKEY_checkpoint_interval_default = "20000" +OCF_RESKEY_compare_timeout_default = "3000" +OCF_RESKEY_expired_scan_cycle_default = "3000" +OCF_RESKEY_use_filter_rewriter_default = "true" +OCF_RESKEY_vnet_hdr_default = "false" +OCF_RESKEY_max_disk_errors_default = "1" +OCF_RESKEY_monitor_timeout_default = "20000" +OCF_RESKEY_yank_timeout_default = "10000" +OCF_RESKEY_fail_fast_timeout_default = "5000" +OCF_RESKEY_debug_default = "0" + +OCF_RESKEY_qemu_binary \ + = os.getenv("OCF_RESKEY_qemu_binary", OCF_RESKEY_qemu_binary_default) +OCF_RESKEY_qemu_img_binary \ + = os.getenv("OCF_RESKEY_qemu_img_binary", OCF_RESKEY_qemu_img_binary_default) +OCF_RESKEY_log_dir \ + = os.getenv("OCF_RESKEY_log_dir", OCF_RESKEY_log_dir_default) +OCF_RESKEY_options \ + = os.getenv("OCF_RESKEY_options", OCF_RESKEY_options_default) +OCF_RESKEY_disk_size \ + = os.getenv("OCF_RESKEY_disk_size", OCF_RESKEY_disk_size_default) +OCF_RESKEY_active_hidden_dir \ + = os.getenv("OCF_RESKEY_active_hidden_dir", OCF_RESKEY_active_hidden_dir_default) +OCF_RESKEY_listen_address \ + = os.getenv("OCF_RESKEY_listen_address", OCF_RESKEY_listen_address_default) +OCF_RESKEY_base_port \ + = os.getenv("OCF_RESKEY_base_port", OCF_RESKEY_base_port_default) +OCF_RESKEY_checkpoint_interval \ + = os.getenv("OCF_RESKEY_checkpoint_interval", OCF_RESKEY_checkpoint_interval_default) +OCF_RESKEY_compare_timeout \ + = os.getenv("OCF_RESKEY_compare_timeout", OCF_RESKEY_compare_timeout_default) +OCF_RESKEY_expired_scan_cycle \ + = os.getenv("OCF_RESKEY_expired_scan_cycle", OCF_RESKEY_expired_scan_cycle_default) +OCF_RESKEY_use_filter_rewriter \ + = os.getenv("OCF_RESKEY_use_filter_rewriter", OCF_RESKEY_use_filter_rewriter_default) +OCF_RESKEY_vnet_hdr \ + = os.getenv("OCF_RESKEY_vnet_hdr", OCF_RESKEY_vnet_hdr_default) +OCF_RESKEY_max_disk_errors \ + = os.getenv("OCF_RESKEY_max_disk_errors", OCF_RESKEY_max_disk_errors_default) +OCF_RESKEY_monitor_timeout \ + = os.getenv("OCF_RESKEY_monitor_timeout", OCF_RESKEY_monitor_timeout_default) +OCF_RESKEY_yank_timeout \ + = os.getenv("OCF_RESKEY_yank_timeout", OCF_RESKEY_yank_timeout_default) +OCF_RESKEY_fail_fast_timeout \ + = os.getenv("OCF_RESKEY_fail_fast_timeout", OCF_RESKEY_fail_fast_timeout_default) +OCF_RESKEY_debug \ + = os.getenv("OCF_RESKEY_debug", OCF_RESKEY_debug_default) + +ACTIVE_IMAGE = os.path.join(OCF_RESKEY_active_hidden_dir, \ + OCF_RESOURCE_INSTANCE + "-active.qcow2") +HIDDEN_IMAGE = os.path.join(OCF_RESKEY_active_hidden_dir, \ + OCF_RESOURCE_INSTANCE + "-hidden.qcow2") + +QMP_SOCK = os.path.join(HA_RSCTMP, OCF_RESOURCE_INSTANCE + "-qmp.sock") +HELPER_SOCK = os.path.join(HA_RSCTMP, OCF_RESOURCE_INSTANCE + "-helper.sock") +COMP_SOCK = os.path.join(HA_RSCTMP, OCF_RESOURCE_INSTANCE + "-compare.sock") +COMP_OUT_SOCK = os.path.join(HA_RSCTMP, OCF_RESOURCE_INSTANCE \ + + "-comp_out.sock") + +PID_FILE = os.path.join(HA_RSCTMP, OCF_RESOURCE_INSTANCE + "-qemu.pid") + +QMP_LOG = os.path.join(OCF_RESKEY_log_dir, OCF_RESOURCE_INSTANCE + "-qmp.log") +QEMU_LOG = os.path.join(OCF_RESKEY_log_dir, OCF_RESOURCE_INSTANCE + "-qemu.log") +HELPER_LOG = os.path.join(OCF_RESKEY_log_dir, OCF_RESOURCE_INSTANCE \ + + "-helper.log") + +START_TIME = time.time() +did_yank = False + +# Exception only raised by ourself +class Error(Exception): + pass + +def setup_constants(): + # This function is called after the parameters where validated + global OCF_RESKEY_CRM_meta_timeout + if OCF_ACTION == "monitor": + OCF_RESKEY_CRM_meta_timeout = OCF_RESKEY_monitor_timeout + + global MIGRATE_PORT, MIRROR_PORT, COMPARE_IN_PORT, NBD_PORT + MIGRATE_PORT = int(OCF_RESKEY_base_port) + MIRROR_PORT = int(OCF_RESKEY_base_port) + 1 + COMPARE_IN_PORT = int(OCF_RESKEY_base_port) + 2 + NBD_PORT = int(OCF_RESKEY_base_port) + 3 + + global QEMU_PRIMARY_CMDLINE + QEMU_PRIMARY_CMDLINE = \ + ("'%(OCF_RESKEY_qemu_binary)s' %(OCF_RESKEY_options)s" + " -drive if=none,node-name=colo-disk0,driver=quorum,read-pattern=fifo," + "vote-threshold=1,children.0=parent0" + " -qmp unix:'%(QMP_SOCK)s',server,nowait" + " -daemonize -D '%(QEMU_LOG)s' -pidfile '%(PID_FILE)s'") % globals() + + global QEMU_SECONDARY_CMDLINE + QEMU_SECONDARY_CMDLINE = \ + ("'%(OCF_RESKEY_qemu_binary)s' %(OCF_RESKEY_options)s" + " -chardev socket,id=red0,host='%(OCF_RESKEY_listen_address)s'," + "port=%(MIRROR_PORT)s,server,nowait,nodelay,yank" + " -chardev socket,id=red1,host='%(OCF_RESKEY_listen_address)s'," + "port=%(COMPARE_IN_PORT)s,server,nowait,nodelay,yank" + " -object filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0" + " -object filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1") \ + % globals() + + if is_true(OCF_RESKEY_use_filter_rewriter): + QEMU_SECONDARY_CMDLINE += \ + " -object filter-rewriter,id=rew0,netdev=hn0,queue=all" + + QEMU_SECONDARY_CMDLINE += \ + (" -drive if=none,node-name=childs0,top-id=colo-disk0," + "driver=replication,mode=secondary,file.driver=qcow2," + "file.file.filename='%(ACTIVE_IMAGE)s',file.backing.driver=qcow2," + "file.backing.file.filename='%(HIDDEN_IMAGE)s'," + "file.backing.backing=parent0" + " -drive if=none,node-name=colo-disk0,driver=quorum,read-pattern=fifo," + "vote-threshold=1,children.0=childs0" + " -incoming tcp:'%(OCF_RESKEY_listen_address)s':%(MIGRATE_PORT)s" + " -global migration.yank=true" + " -qmp unix:'%(QMP_SOCK)s',server,nowait" + " -daemonize -D '%(QEMU_LOG)s' -pidfile '%(PID_FILE)s'") % globals() + +def qemu_colo_meta_data(): + print("""\ + + + + + 1.0 + +Resource agent for qemu COLO. (https://wiki.qemu.org/Features/COLO) + +After defining the master/slave instance, the master score has to be +manually set to show which node has up-to-date data. So you copy your +image to one host (and create empty images the other host(s)) and then +run "crm_master -r name_of_your_primitive -v 10" on that host. +Also, you have to set 'notify=true' in the metadata attributes when +defining the master/slave instance. + +Note: +-If the instance is stopped cluster-wide, the resource agent will do a +clean shutdown. Set the demote timeout to the time it takes for your +guest to shutdown. +-Colo replication is started from the monitor action. Set the monitor +timeout to at least the time it takes for replication to start. You can +set the monitor_timeout parameter for a soft timeout, which the resource +agent tries to satisfy. +-The resource agent may notify pacemaker about peer failure, +these failures will show up with exitreason="Simulated failure". + + Qemu COLO + + + + + qemu binary to use + qemu binary + + + + + qemu-img binary to use + qemu-img binary + + + + + Directory to place logs in + Log directory + + + + + +Options to pass to qemu. These will be passed alongside COLO specific +options, so you need to follow these conventions: The netdev should have +id=hn0 and the disk controller drive=colo-disk0. The image node should +have id/node-name=parent0, but should not be connected to the guest. +Example: +-vnc :0 -enable-kvm -cpu qemu64,+kvmclock -m 512 -netdev bridge,id=hn0 +-device e1000,netdev=hn0 -device virtio-blk,drive=colo-disk0 +-drive if=none,id=parent0,format=qcow2,file=/mnt/vms/vm01.qcow2 + + Options to pass to qemu. + + + + Disk size of the image + Disk size of the image + + + + + +Directory where the active and hidden images will be stored. It is +recommended to put this on a ramdisk. + + Path to active and hidden images + + + + + Address to listen on. + Listen address + + + + + +4 tcp ports that are unique for each instance. (base_port to base_port + 3) + + Ports to use + + + + + +Interval for regular checkpoints in milliseconds. + + Interval for regular checkpoints + + + + + +Maximum time to hold a primary packet if secondary hasn't sent it yet, +in milliseconds. +You should also adjust "expired_scan_cycle" accordingly. + + Compare timeout + + + + + +Interval for checking for expired primary packets in milliseconds. + + Expired packet check interval + + + + + +Use filter-rewriter to increase similarity between the VMs. + + Use filter-rewriter + + + + + +Set this to true if your system supports vnet_hdr and you enabled +it on the tap netdev. + + vnet_hdr support + + + + + +Maximum disk read errors per monitor interval before marking the resource +as failed. A write error is always fatal except if the value is 0. +A value of 0 will disable disk error handling. +Primary disk errors are only handled if there is a healthy secondary. + + Maximum disk errors + + + + + +Soft timeout for monitor, in milliseconds. +Must be lower than the monitor action timeout. + + Monitor timeout + + + + + +Timeout for QMP commands after which to execute the "yank" command, +in milliseconds. +Must be lower than any of the action timeouts. + + Yank timeout + + + + + +Timeout for QMP commands used in the stop and demote actions to speed +up recovery from a hanging qemu, in milliseconds. +Must be lower than any of the action timeouts. + + Timeout for fast paths + + + + + +Control debugging: +0: disable debugging +1: log debug messages and qmp commands +2: + dump core of hanging qemu + + Control debugging + + + + + + + + + + + + + + + + + + +""") + +def logs_open(): + global log + log = logging.getLogger(OCF_RESOURCE_INSTANCE) + if int(OCF_RESKEY_debug) >= 1 or HA_DEBUG != "0": + log.setLevel(logging.DEBUG) + else: + log.setLevel(logging.INFO) + + formater = logging.Formatter("(%(name)s) %(levelname)s: %(message)s") + + if sys.stdout.isatty(): + handler = logging.StreamHandler(stream=sys.stderr) + handler.setFormatter(formater) + log.addHandler(handler) + + if HA_LOGFACILITY: + handler = logging.handlers.SysLogHandler("/dev/log") + handler.setFormatter(formater) + log.addHandler(handler) + + if HA_LOGFILE: + handler = logging.FileHandler(HA_LOGFILE) + handler.setFormatter(formater) + log.addHandler(handler) + + if HA_DEBUGLOG and HA_DEBUGLOG != HA_LOGFILE: + handler = logging.FileHandler(HA_DEBUGLOG) + handler.setFormatter(formater) + log.addHandler(handler) + + global qmp_log + qmp_log = logging.getLogger("qmp_log") + qmp_log.setLevel(logging.DEBUG) + formater = logging.Formatter("%(message)s") + + if int(OCF_RESKEY_debug) >= 1: + handler = logging.handlers.WatchedFileHandler(QMP_LOG) + handler.setFormatter(formater) + qmp_log.addHandler(handler) + else: + handler = logging.NullHandler() + qmp_log.addHandler(handler) + +def rotate_logfile(logfile, numlogs): + numlogs -= 1 + for n in range(numlogs, -1, -1): + file = logfile + if n != 0: + file = "%s.%s" % (file, n) + if os.path.exists(file): + if n == numlogs: + os.remove(file) + else: + newname = "%s.%s" % (logfile, n + 1) + os.rename(file, newname) + +def is_writable(file): + return os.access(file, os.W_OK) + +def is_executable_file(file): + return os.path.isfile(file) and os.access(file, os.X_OK) + +def is_true(var): + return re.match("yes|true|1|YES|TRUE|True|ja|on|ON", str(var)) != None + +# Check if the binary exists and is executable +def check_binary(binary): + if is_executable_file(binary): + return True + PATH = os.getenv("PATH", os.defpath) + for dir in PATH.split(os.pathsep): + if is_executable_file(os.path.join(dir, binary)): + return True + log.error("binary \"%s\" doesn't exist or not executable" % binary) + return False + +def run_command(commandline): + proc = subprocess.Popen(commandline, shell=True, stdout=subprocess.PIPE, + stderr=subprocess.STDOUT, universal_newlines=True) + stdout, stderr = proc.communicate() + if proc.returncode != 0: + log.error("command \"%s\" failed with code %s:\n%s" \ + % (commandline, proc.returncode, stdout)) + raise Error() + +# Functions for setting and getting the master score to tell Pacemaker which +# host has the most recent data +def set_master_score(score): + if score == 0: + run_command("crm_master -q -l forever -D") + else: + run_command("crm_master -q -l forever -v %s" % score) + +def set_remote_master_score(remote, score): + if score == 0: + run_command("crm_master -q -l forever -N '%s' -D" % remote) + else: + run_command("crm_master -q -l forever -N '%s' -v %s" % (remote, score)) + +def get_master_score(): + proc = subprocess.Popen("crm_master -q -G", shell=True, + stdout=subprocess.PIPE, stderr=subprocess.DEVNULL, + universal_newlines=True) + stdout, stderr = proc.communicate() + if proc.returncode != 0: + return 0 + else: + return int(str.strip(stdout)) + +def get_remote_master_score(remote): + proc = subprocess.Popen("crm_master -q -N '%s' -G" % remote, shell=True, + stdout=subprocess.PIPE, stderr=subprocess.DEVNULL, + universal_newlines=True) + stdout, stderr = proc.communicate() + if proc.returncode != 0: + return 0 + else: + return int(str.strip(stdout)) + +# Tell Pacemaker that the remote resource failed +def report_remote_failure(remote): + run_command("crm_resource --resource '%s' --fail --node '%s'" + % (OCF_RESOURCE_INSTANCE, remote)) + +def recv_line(fd): + line = "" + while True: + tmp = fd.recv(1).decode() + line += tmp + if tmp == "\n" or len(tmp) == 0: + break + return line + +# Filter out events +def read_answer(fd): + while True: + line = recv_line(fd) + qmp_log.debug(str.strip(line)) + + if len(line) == 0: + log.error("qmp connection closed") + raise Error() + + answer = json.loads(line) + # Ignore everything else + if "return" in answer or "error" in answer: + break + return answer + +# Execute one or more qmp commands +def qmp_execute(fd, commands, ignore_error = False, do_yank = True): + for command in commands: + if not command: + continue + + try: + to_send = json.dumps(command) + fd.sendall(str.encode(to_send + "\n")) + qmp_log.debug(to_send) + + answer = read_answer(fd) + except Exception as e: + if isinstance(e, socket.timeout) and do_yank: + log.warning("Command timed out, trying to unfreeze qemu") + new_timeout = max(2, (int(OCF_RESKEY_CRM_meta_timeout)/1000) \ + - (time.time() - START_TIME) - 2) + fd.settimeout(new_timeout) + answer = yank(fd) + # Read answer of timed-out command + try: + if "id" in answer: + answer = read_answer(fd) + else: + read_answer(fd) + except socket.error as e: + log.error("while reading answer of timed out command: " + "%s\n%s" % (json.dumps(command), e)) + raise Error() + elif isinstance(e, socket.timeout) or isinstance(e, socket.error): + log.error("while executing qmp command: %s\n%s" \ + % (json.dumps(command), e)) + raise Error() + else: + raise + + if not ignore_error and ("error" in answer): + log.error("qmp command returned error:\n%s\n%s" \ + % (json.dumps(command), json.dumps(answer))) + raise Error() + + return answer + +# Open qemu qmp connection +def qmp_open(fail_fast = False): + if fail_fast: + timeout = int(OCF_RESKEY_fail_fast_timeout)/1000 + else: + timeout = int(OCF_RESKEY_yank_timeout)/1000 + + try: + fd = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) + fd.settimeout(timeout) + fd.connect(HELPER_SOCK) + except socket.error as e: + log.error("while connecting to helper socket: %s" % e) + raise Error() + + return fd + +def yank(fd): + global did_yank + did_yank = True + answer = qmp_execute(fd, [{"exec-oob": "yank", "id": "yank0"}], \ + do_yank = False, ignore_error = True) + return answer + +def oob_helper_exec(client, cmd, events): + if cmd["exec-helper"] == "get-events": + event = cmd["arguments"]["event"] + if (event in events): + to_send = json.dumps({"return": events[event]}) + client.sendall(str.encode(to_send + "\n")) + else: + client.sendall(str.encode("{\"return\": []}\n")) + elif cmd["exec-helper"] == "clear-events": + events.clear() + client.sendall(str.encode("{\"return\": {}}\n")) + else: + client.sendall(str.encode("{\"error\": \"Unknown helper command\"}\n")) + +def oob_helper(qmp, server): + max_events = max(100, int(OCF_RESKEY_max_disk_errors)) + events = {} + try: + os.close(0) + os.close(1) + os.close(2) + logging.shutdown() + + client = None + while True: + if client: + watch = [client, qmp] + else: + watch = [server, qmp] + sel = select.select(watch, [], []) + try: + if client in sel[0]: + cmd = recv_line(client) + if len(cmd) == 0: + # client socket was closed: wait for new client + client.close() + client = None + continue + else: + parsed = json.loads(cmd) + if ("exec-helper" in parsed): + oob_helper_exec(client, parsed, events) + else: + qmp.sendall(str.encode(cmd)) + if qmp in sel[0]: + answer = recv_line(qmp) + if len(answer) == 0: + # qmp socket was closed: qemu died, exit + os._exit(0) + else: + parsed = json.loads(answer) + if ("event" in parsed): + event = parsed["event"] + if (event not in events): + events[event] = [] + if len(events[event]) < max_events: + events[event].append(parsed) + elif client: + client.sendall(str.encode(answer)) + if server in sel[0]: + client, client_addr = server.accept() + except socket.error as e: + pass + except Exception as e: + with open(HELPER_LOG, 'a') as f: + f.write(str(e) + "\n") + os._exit(0) + +# Fork off helper to keep the oob qmp connection open and to catch events +def oob_helper_open(): + try: + qmp = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) + qmp.connect(QMP_SOCK) + qmp_execute(qmp, [{"execute": "qmp_capabilities", "arguments": {"enable": ["oob"]}}]) + except socket.error as e: + log.error("while connecting to qmp socket: %s" % e) + raise Error() + + try: + if os.path.exists(HELPER_SOCK): + os.unlink(HELPER_SOCK) + server = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) + server.bind(HELPER_SOCK) + server.listen(1) + except socket.error as e: + log.error("while opening helper socket: %s" % e) + raise Error() + + qmp.set_inheritable(True) + server.set_inheritable(True) + + try: + pid = os.fork() + except OSError as e: + log.error("while forking off oob helper: %s" % e) + raise Error() + + if pid == 0: + # child 1: Exits after forking off child 2, so pid 1 will become + # responsible for the child + os.setsid() + + pid = os.fork() + + if pid == 0: + # child 2: here the actual work is being done + oob_helper(qmp, server) + else: + os._exit(0) + + qmp.close() + server.close() + +# Get the host of the nbd node +def qmp_get_nbd_remote(fd): + block_nodes = qmp_execute(fd, [{"execute": "query-named-block-nodes", "arguments": {"flat": True}}]) + for node in block_nodes["return"]: + if node["node-name"] == "nbd0": + url = str(node["image"]["filename"]) + return str.split(url, "//")[1].split("/")[0].split(":")[0] + return None + +# Check if we are currently resyncing +def qmp_check_resync(fd): + answer = qmp_execute(fd, [{"execute": "query-block-jobs"}]) + for job in answer["return"]: + if job["device"] == "resync": + return job + return None + +def qmp_start_resync(fd, remote): + answer = qmp_execute(fd, [{"execute": "blockdev-add", "arguments": {"driver": "nbd", "node-name": "nbd0", "server": {"type": "inet", "host": str(remote), "port": str(NBD_PORT)}, "export": "parent0", "detect-zeroes": "on", "yank": True}}], ignore_error = True) + if "error" in answer: + log.warning("Failed to add nbd node: %s" % json.dumps(answer)) + log.warning("Assuming peer failure") + report_remote_failure(remote) + else: + qmp_execute(fd, [{"execute": "blockdev-mirror", "arguments": {"device": "colo-disk0", "job-id": "resync", "target": "nbd0", "sync": "full", "on-target-error": "report", "on-source-error": "ignore", "auto-dismiss": False}}]) + +def qmp_cancel_resync(fd): + timeout = START_TIME + (int(OCF_RESKEY_yank_timeout)/1000) + + if qmp_check_resync(fd)["status"] != "concluded": + qmp_execute(fd, [{"execute": "block-job-cancel", "arguments": {"device": "resync", "force": True}}], ignore_error = True) + # Wait for the block-job to finish + while time.time() < timeout: + if qmp_check_resync(fd)["status"] == "concluded": + break + log.debug("Waiting for block-job to finish in qmp_cancel_resync()") + time.sleep(1) + else: + log.warning("Timed out, trying to unfreeze qemu") + yank(fd) + while qmp_check_resync(fd)["status"] != "concluded": + log.debug("Waiting for block-job to finish") + time.sleep(1) + + qmp_execute(fd, [ + {"execute": "block-job-dismiss", "arguments": {"id": "resync"}}, + {"execute": "blockdev-del", "arguments": {"node-name": "nbd0"}} + ]) + +def qmp_start_colo(fd, remote): + # Check if we have a filter-rewriter + answer = qmp_execute(fd, [{"execute": "qom-list", "arguments": {"path": "/objects/rew0"}}], ignore_error = True) + if "error" in answer: + if answer["error"]["class"] == "DeviceNotFound": + have_filter_rewriter = False + else: + log.error("while checking for filter-rewriter:\n%s" \ + % json.dumps(answer)) + raise Error() + else: + have_filter_rewriter = True + + # Pause VM and cancel resync + qmp_execute(fd, [ + {"execute": "stop"}, + {"execute": "block-job-cancel", "arguments": {"device": "resync"}} + ]) + + # Wait for the block-job to finish + while qmp_check_resync(fd)["status"] != "concluded": + log.debug("Waiting for block-job to finish in qmp_start_colo()") + time.sleep(1) + + # Add nbd to the quorum node + qmp_execute(fd, [ + {"execute": "block-job-dismiss", "arguments": {"id": "resync"}}, + {"execute": "x-blockdev-change", "arguments": {"parent": "colo-disk0", "node": "nbd0"}} + ]) + + # Connect mirror and compare_in to secondary + qmp_execute(fd, [ + {"execute": "chardev-add", "arguments": {"id": "comp_pri_in0<", "backend": {"type": "socket", "data": {"addr": {"type": "unix", "data": {"path": str(COMP_SOCK)}}, "server": True}}}}, + {"execute": "chardev-add", "arguments": {"id": "comp_pri_in0>", "backend": {"type": "socket", "data": {"addr": {"type": "unix", "data": {"path": str(COMP_SOCK)}}, "server": False}}}}, + {"execute": "chardev-add", "arguments": {"id": "comp_out0<", "backend": {"type": "socket", "data": {"addr": {"type": "unix", "data": {"path": str(COMP_OUT_SOCK)}}, "server": True}}}}, + {"execute": "chardev-add", "arguments": {"id": "comp_out0>", "backend": {"type": "socket", "data": {"addr": {"type": "unix", "data": {"path": str(COMP_OUT_SOCK)}}, "server": False}}}}, + {"execute": "chardev-add", "arguments": {"id": "mirror0", "backend": {"type": "socket", "data": {"addr": {"type": "inet", "data": {"host": str(remote), "port": str(MIRROR_PORT)}}, "server": False, "nodelay": True, "yank": True}}}}, + {"execute": "chardev-add", "arguments": {"id": "comp_sec_in0", "backend": {"type": "socket", "data": {"addr": {"type": "inet", "data": {"host": str(remote), "port": str(COMPARE_IN_PORT)}}, "server": False, "nodelay": True, "yank": True}}}} + ]) + + # Add the COLO filters + vnet_hdr_support = is_true(OCF_RESKEY_vnet_hdr) + if have_filter_rewriter: + qmp_execute(fd, [ + {"execute": "object-add", "arguments": {"qom-type": "filter-mirror", "id": "m0", "props": {"insert": "before", "position": "id=rew0", "netdev": "hn0", "queue": "tx", "outdev": "mirror0", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "filter-redirector", "id": "redire0", "props": {"insert": "before", "position": "id=rew0", "netdev": "hn0", "queue": "rx", "indev": "comp_out0<", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "filter-redirector", "id": "redire1", "props": {"insert": "before", "position": "id=rew0", "netdev": "hn0", "queue": "rx", "outdev": "comp_pri_in0<", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "iothread", "id": "iothread1"}}, + {"execute": "object-add", "arguments": {"qom-type": "colo-compare", "id": "comp0", "props": {"primary_in": "comp_pri_in0>", "secondary_in": "comp_sec_in0", "outdev": "comp_out0>", "iothread": "iothread1", "compare_timeout": int(OCF_RESKEY_compare_timeout), "expired_scan_cycle": int(OCF_RESKEY_expired_scan_cycle), "vnet_hdr_support": vnet_hdr_support}}} + ]) + else: + qmp_execute(fd, [ + {"execute": "object-add", "arguments": {"qom-type": "filter-mirror", "id": "m0", "props": {"netdev": "hn0", "queue": "tx", "outdev": "mirror0", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "filter-redirector", "id": "redire0", "props": {"netdev": "hn0", "queue": "rx", "indev": "comp_out0<", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "filter-redirector", "id": "redire1", "props": {"netdev": "hn0", "queue": "rx", "outdev": "comp_pri_in0<", "vnet_hdr_support": vnet_hdr_support}}}, + {"execute": "object-add", "arguments": {"qom-type": "iothread", "id": "iothread1"}}, + {"execute": "object-add", "arguments": {"qom-type": "colo-compare", "id": "comp0", "props": {"primary_in": "comp_pri_in0>", "secondary_in": "comp_sec_in0", "outdev": "comp_out0>", "iothread": "iothread1", "compare_timeout": int(OCF_RESKEY_compare_timeout), "expired_scan_cycle": int(OCF_RESKEY_expired_scan_cycle), "vnet_hdr_support": vnet_hdr_support}}} + ]) + + # Start COLO + qmp_execute(fd, [ + {"execute": "migrate-set-capabilities", "arguments": {"capabilities": [{"capability": "x-colo", "state": True }] }}, + {"execute": "migrate-set-parameters" , "arguments": {"x-checkpoint-delay": int(OCF_RESKEY_checkpoint_interval), "yank": True}}, + {"execute": "migrate", "arguments": {"uri": "tcp:%s:%s" % (remote, MIGRATE_PORT)}} + ]) + + # Wait for COLO to start + while qmp_execute(fd, [{"execute": "query-status"}])["return"]["status"] \ + == "paused" \ + or qmp_execute(fd, [{"execute": "query-colo-status"}])["return"]["mode"] \ + != "primary" : + log.debug("Waiting for colo replication to start") + time.sleep(1) + +def qmp_primary_failover(fd): + qmp_execute(fd, [ + {"execute": "object-del", "arguments": {"id": "m0"}}, + {"execute": "object-del", "arguments": {"id": "redire0"}}, + {"execute": "object-del", "arguments": {"id": "redire1"}}, + {"execute": "x-colo-lost-heartbeat"}, + {"execute": "object-del", "arguments": {"id": "comp0"}}, + {"execute": "object-del", "arguments": {"id": "iothread1"}}, + {"execute": "x-blockdev-change", "arguments": {"parent": "colo-disk0", "child": "children.1"}}, + {"execute": "blockdev-del", "arguments": {"node-name": "nbd0"}}, + {"execute": "chardev-remove", "arguments": {"id": "mirror0"}}, + {"execute": "chardev-remove", "arguments": {"id": "comp_sec_in0"}}, + {"execute": "chardev-remove", "arguments": {"id": "comp_pri_in0>"}}, + {"execute": "chardev-remove", "arguments": {"id": "comp_pri_in0<"}}, + {"execute": "chardev-remove", "arguments": {"id": "comp_out0>"}}, + {"execute": "chardev-remove", "arguments": {"id": "comp_out0<"}} + ]) + +def qmp_secondary_failover(fd): + qmp_execute(fd, [ + {"execute": "nbd-server-stop"}, + {"execute": "object-del", "arguments": {"id": "f2"}}, + {"execute": "object-del", "arguments": {"id": "f1"}}, + {"execute": "x-colo-lost-heartbeat"}, + {"execute": "chardev-remove", "arguments": {"id": "red1"}}, + {"execute": "chardev-remove", "arguments": {"id": "red0"}}, + ]) + +# Check qemu health and colo role +def qmp_check_state(fd, fail_fast = False): + answer = qmp_execute(fd, [{"execute": "query-status"}], \ + do_yank = not fail_fast) + vm_status = answer["return"] + + answer = qmp_execute(fd, [{"execute": "query-colo-status"}], \ + do_yank = not fail_fast) + colo_status = answer["return"] + + if vm_status["status"] == "inmigrate": + role = OCF_SUCCESS + replication = OCF_NOT_RUNNING + + elif (vm_status["status"] == "running" \ + or vm_status["status"] == "colo" \ + or vm_status["status"] == "finish-migrate") \ + and colo_status["mode"] == "none" \ + and (colo_status["reason"] == "request" \ + or colo_status["reason"] == "none"): + role = OCF_RUNNING_MASTER + replication = OCF_NOT_RUNNING + + elif (vm_status["status"] == "running" \ + or vm_status["status"] == "colo" \ + or vm_status["status"] == "finish-migrate") \ + and colo_status["mode"] == "secondary": + role = OCF_SUCCESS + replication = OCF_SUCCESS + + elif (vm_status["status"] == "running" \ + or vm_status["status"] == "colo" \ + or vm_status["status"] == "finish-migrate") \ + and colo_status["mode"] == "primary": + role = OCF_RUNNING_MASTER + replication = OCF_SUCCESS + + else: + log.error("Invalid qemu status:\nvm status: %s\ncolo status: %s" \ + % (vm_status, colo_status)) + role = OCF_ERR_GENERIC + replication = OCF_ERR_GENERIC + + return role, replication + +# Sanity checks: check parameters, files, binaries, etc. +def qemu_colo_validate_all(): + # Check resource parameters + if not str.isdigit(OCF_RESKEY_base_port): + log.error("base_port needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_checkpoint_interval): + log.error("checkpoint_interval needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_compare_timeout): + log.error("compare_timeout needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_expired_scan_cycle): + log.error("expired_scan_cycle needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_max_disk_errors): + log.error("max_disk_errors needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_monitor_timeout): + log.error("monitor_timeout needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_yank_timeout): + log.error("yank_timeout needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_fail_fast_timeout): + log.error("fail_fast_timeout needs to be a number") + return OCF_ERR_CONFIGURED + + if not str.isdigit(OCF_RESKEY_debug): + log.error("debug needs to be a number") + return OCF_ERR_CONFIGURED + + if not OCF_RESKEY_active_hidden_dir: + log.error("active_hidden_dir needs to be specified") + return OCF_ERR_CONFIGURED + + if not OCF_RESKEY_disk_size: + log.error("disk_size needs to be specified") + return OCF_ERR_CONFIGURED + + # Check resource meta configuration + if OCF_ACTION != "stop": + if OCF_RESKEY_CRM_meta_master_max != 1: + log.error("only one master allowed") + return OCF_ERR_CONFIGURED + + if OCF_RESKEY_CRM_meta_clone_max > 2: + log.error("maximum 2 clones allowed") + return OCF_ERR_CONFIGURED + + if OCF_RESKEY_CRM_meta_master_node_max != 1: + log.error("only one master per node allowed") + return OCF_ERR_CONFIGURED + + if OCF_RESKEY_CRM_meta_clone_node_max != 1: + log.error("only one clone per node allowed") + return OCF_ERR_CONFIGURED + + # Check if notify is enabled + if OCF_ACTION != "stop" and OCF_ACTION != "monitor": + if not is_true(OCF_RESKEY_CRM_meta_notify) \ + and not OCF_RESKEY_CRM_meta_notify_start_uname: + log.error("notify needs to be enabled") + return OCF_ERR_CONFIGURED + + # Check that globally-unique is disabled + if is_true(OCF_RESKEY_CRM_meta_globally_unique): + log.error("globally-unique needs to be disabled") + return OCF_ERR_CONFIGURED + + # Check binaries + if not check_binary(OCF_RESKEY_qemu_binary): + return OCF_ERR_INSTALLED + + if not check_binary(OCF_RESKEY_qemu_img_binary): + return OCF_ERR_INSTALLED + + # Check paths and files + if not is_writable(OCF_RESKEY_active_hidden_dir) \ + or not os.path.isdir(OCF_RESKEY_active_hidden_dir): + log.error("active and hidden image directory missing or not writable") + return OCF_ERR_PERM + + return OCF_SUCCESS + +# Check if qemu is running +def check_pid(): + if not os.path.exists(PID_FILE): + return OCF_NOT_RUNNING, None + + fd = open(PID_FILE, "r") + pid = int(str.strip(fd.readline())) + fd.close() + try: + os.kill(pid, 0) + except OSError: + log.info("qemu is not running") + return OCF_NOT_RUNNING, pid + else: + return OCF_SUCCESS, pid + +def qemu_colo_monitor(fail_fast = False): + status, pid = check_pid() + if status != OCF_SUCCESS: + return status, OCF_NOT_RUNNING + + fd = qmp_open(fail_fast) + + role, replication = qmp_check_state(fd, fail_fast) + if role != OCF_SUCCESS and role != OCF_RUNNING_MASTER: + return role, replication + + colo_events = qmp_execute(fd, [{"exec-helper": "get-events", "arguments": {"event": "COLO_EXIT"}}], do_yank = False) + for event in colo_events["return"]: + if event["data"]["reason"] == "error": + if replication == OCF_SUCCESS: + replication = OCF_ERR_GENERIC + + if did_yank and replication == OCF_SUCCESS: + replication = OCF_ERR_GENERIC + + peer_disk_errors = 0 + local_disk_errors = 0 + quorum_events = qmp_execute(fd, [{"exec-helper": "get-events", "arguments": {"event": "QUORUM_REPORT_BAD"}}], do_yank = False) + for event in quorum_events["return"]: + if event["data"]["node-name"] == "nbd0": + if event["data"]["type"] == "read": + peer_disk_errors += 1 + else: + peer_disk_errors += int(OCF_RESKEY_max_disk_errors) + else: + if event["data"]["type"] == "read": + local_disk_errors += 1 + else: + local_disk_errors += int(OCF_RESKEY_max_disk_errors) + + if int(OCF_RESKEY_max_disk_errors) != 0: + if peer_disk_errors >= int(OCF_RESKEY_max_disk_errors): + log.error("Peer disk error") + if replication == OCF_SUCCESS: + replication = OCF_ERR_GENERIC + + if local_disk_errors >= int(OCF_RESKEY_max_disk_errors): + if replication == OCF_SUCCESS: + log.error("Local disk error") + role = OCF_ERR_GENERIC + else: + log.warning("Local disk error") + + if not fail_fast and OCF_RESKEY_CRM_meta_interval != 0: + # This isn't a probe monitor + block_job = qmp_check_resync(fd) + if block_job: + if "error" in block_job: + log.error("resync error: %s" % block_job["error"]) + peer = qmp_get_nbd_remote(fd) + qmp_cancel_resync(fd) + report_remote_failure(peer) + elif block_job["ready"] == True: + log.info("resync done, starting colo") + peer = qmp_get_nbd_remote(fd) + qmp_start_colo(fd, peer) + # COLO started, our secondary now can be promoted if the + # primary fails + set_remote_master_score(peer, 100) + else: + pct_done = (float(block_job["offset"]) \ + / float(block_job["len"])) * 100 + log.info("resync %.1f%% done" % pct_done) + else: + if replication == OCF_ERR_GENERIC: + if role == OCF_RUNNING_MASTER: + log.error("Replication error") + peer = qmp_get_nbd_remote(fd) + if peer: + report_remote_failure(peer) + else: + log.warning("Replication error") + qmp_execute(fd, [{"exec-helper": "clear-events"}], do_yank = False) + + fd.close() + + return role, replication + +def qemu_colo_start(): + if check_pid()[0] == OCF_SUCCESS: + log.info("qemu is already running") + return OCF_SUCCESS + + run_command("'%s' create -q -f qcow2 %s %s" \ + % (OCF_RESKEY_qemu_img_binary, ACTIVE_IMAGE, OCF_RESKEY_disk_size)) + run_command("'%s' create -q -f qcow2 %s %s" \ + % (OCF_RESKEY_qemu_img_binary, HIDDEN_IMAGE, OCF_RESKEY_disk_size)) + + rotate_logfile(QMP_LOG, 8) + rotate_logfile(QEMU_LOG, 8) + run_command(QEMU_SECONDARY_CMDLINE) + oob_helper_open() + + fd = qmp_open() + qmp_execute(fd, [ + {"execute": "nbd-server-start", "arguments": {"addr": {"type": "inet", "data": {"host": str(OCF_RESKEY_listen_address), "port": str(NBD_PORT)}}}}, + {"execute": "nbd-server-add", "arguments": {"device": "parent0", "writable": True}} + ]) + fd.close() + + return OCF_SUCCESS + +def env_do_shutdown_guest(): + return OCF_RESKEY_CRM_meta_notify_active_uname \ + and OCF_RESKEY_CRM_meta_notify_stop_uname \ + and str.strip(OCF_RESKEY_CRM_meta_notify_active_uname) \ + == str.strip(OCF_RESKEY_CRM_meta_notify_stop_uname) + +def env_find_secondary(): + # slave(s) = + # OCF_RESKEY_CRM_meta_notify_slave_uname + # - OCF_RESKEY_CRM_meta_notify_stop_uname + # + OCF_RESKEY_CRM_meta_notify_start_uname + # Filter out hosts that are stopping and ourselves + for host in str.split(OCF_RESKEY_CRM_meta_notify_slave_uname, " "): + if host: + for stopping_host \ + in str.split(OCF_RESKEY_CRM_meta_notify_stop_uname, " "): + if host == stopping_host: + break + else: + if host != HOSTNAME: + # we found a valid secondary + return host + + for host in str.split(OCF_RESKEY_CRM_meta_notify_start_uname, " "): + if host != HOSTNAME: + # we found a valid secondary + return host + + # we found no secondary + return None + +def _qemu_colo_stop(monstatus, shutdown_guest): + # stop action must do everything possible to stop the resource + try: + timeout = START_TIME + (int(OCF_RESKEY_CRM_meta_timeout)/1000) - 5 + force_stop = False + + if monstatus == OCF_NOT_RUNNING: + log.info("resource is already stopped") + return OCF_SUCCESS + elif monstatus == OCF_RUNNING_MASTER or monstatus == OCF_SUCCESS: + force_stop = False + else: + force_stop = True + + if not force_stop: + fd = qmp_open(True) + if shutdown_guest: + if monstatus == OCF_RUNNING_MASTER: + qmp_execute(fd, [{"execute": "system_powerdown"}], \ + do_yank = False) + else: + qmp_execute(fd, [{"execute": "quit"}], do_yank = False) + fd.close() + + # wait for qemu to stop + while time.time() < timeout: + status, pid = check_pid() + if status == OCF_NOT_RUNNING: + # qemu stopped + return OCF_SUCCESS + elif status == OCF_SUCCESS: + # wait + log.debug("Waiting for qemu to stop") + time.sleep(1) + else: + # something went wrong, force stop instead + break + + log.warning("clean stop timeout reached") + except Exception as e: + log.warning("error while stopping: %s" % e) + + log.info("force stopping qemu") + + status, pid = check_pid() + if status == OCF_NOT_RUNNING: + return OCF_SUCCESS + try: + if int(OCF_RESKEY_debug) >= 2: + os.kill(pid, signal.SIGSEGV) + else: + os.kill(pid, signal.SIGTERM) + time.sleep(2) + os.kill(pid, signal.SIGKILL) + except Exception: + pass + + while check_pid()[0] != OCF_NOT_RUNNING: + time.sleep(1) + + return OCF_SUCCESS + +def qemu_colo_stop(): + shutdown_guest = env_do_shutdown_guest() + try: + role, replication = qemu_colo_monitor(True) + except Exception: + role, replication = OCF_ERR_GENERIC, OCF_ERR_GENERIC + + status = _qemu_colo_stop(role, shutdown_guest) + + if HOSTNAME == str.strip(OCF_RESKEY_CRM_meta_notify_master_uname): + if str.strip(OCF_RESKEY_CRM_meta_notify_promote_uname) != HOSTNAME: + # We where primary and the secondary is to be promoted. + # We are going to be out of date. + set_master_score(0) + else: + if role == OCF_RUNNING_MASTER: + # We where a healthy primary but had no healty secondary or it + # was stopped as well. So we have up-to-date data. + set_master_score(10) + else: + # We where a unhealthy primary but also had no healty secondary. + # So we still should have up-to-date data. + set_master_score(5) + else: + if get_master_score() > 10: + if role == OCF_SUCCESS: + if shutdown_guest: + # We where a healthy secondary and (probably) had a healthy + # primary and both where stopped. So we have up-to-date data + # too. + set_master_score(10) + else: + # We where a healthy secondary and (probably) had a healthy + # primary still running. So we are now out of date. + set_master_score(0) + else: + # We where a unhealthy secondary. So we are now out of date. + set_master_score(0) + + return status + +def qemu_colo_notify(): + action = "%s-%s" % (OCF_RESKEY_CRM_meta_notify_type, \ + OCF_RESKEY_CRM_meta_notify_operation) + + if action == "post-start": + if HOSTNAME == str.strip(OCF_RESKEY_CRM_meta_notify_master_uname): + peer = str.strip(OCF_RESKEY_CRM_meta_notify_start_uname) + fd = qmp_open() + qmp_start_resync(fd, peer) + # The secondary has inconsistent data until resync is finished + set_remote_master_score(peer, 0) + fd.close() + + elif action == "pre-stop": + if not env_do_shutdown_guest() \ + and HOSTNAME == str.strip(OCF_RESKEY_CRM_meta_notify_master_uname) \ + and HOSTNAME != str.strip(OCF_RESKEY_CRM_meta_notify_stop_uname): + fd = qmp_open() + peer = qmp_get_nbd_remote(fd) + log.debug("our peer: %s" % peer) + if peer == str.strip(OCF_RESKEY_CRM_meta_notify_stop_uname): + if qmp_check_resync(fd): + qmp_cancel_resync(fd) + elif get_remote_master_score(peer) > 10: + qmp_primary_failover(fd) + qmp_execute(fd, [{"exec-helper": "clear-events"}],do_yank=False) + fd.close() + + elif action == "post-stop" \ + and OCF_RESKEY_CRM_meta_notify_key_operation == "stonith" \ + and (HOSTNAME == str.strip(OCF_RESKEY_CRM_meta_notify_master_uname) + or str.strip(OCF_RESKEY_CRM_meta_notify_promote_uname)): + peer = str.strip(OCF_RESKEY_CRM_meta_notify_stop_uname) + set_remote_master_score(peer, 0) + + return OCF_SUCCESS + +def qemu_colo_promote(): + role, replication = qemu_colo_monitor() + + if role == OCF_SUCCESS and replication == OCF_NOT_RUNNING: + status = _qemu_colo_stop(role, False) + if status != OCF_SUCCESS: + return status + + rotate_logfile(QMP_LOG, 8) + rotate_logfile(QEMU_LOG, 8) + run_command(QEMU_PRIMARY_CMDLINE) + oob_helper_open() + set_master_score(101) + + peer = env_find_secondary() + if peer: + fd = qmp_open() + qmp_start_resync(fd, peer) + # The secondary has inconsistent data until resync is finished + set_remote_master_score(peer, 0) + fd.close() + return OCF_SUCCESS + elif role == OCF_SUCCESS and replication != OCF_NOT_RUNNING: + fd = qmp_open() + qmp_secondary_failover(fd) + set_master_score(101) + + peer = env_find_secondary() + if peer: + qmp_start_resync(fd, peer) + # The secondary has inconsistent data until resync is finished + set_remote_master_score(peer, 0) + qmp_execute(fd, [{"exec-helper": "clear-events"}], do_yank=False) + fd.close() + return OCF_SUCCESS + else: + return OCF_ERR_GENERIC + +def qemu_colo_demote(): + status = qemu_colo_stop() + if status != OCF_SUCCESS: + return status + return qemu_colo_start() + + +if OCF_ACTION == "meta-data": + qemu_colo_meta_data() + exit(OCF_SUCCESS) + +logs_open() + +status = qemu_colo_validate_all() +# Exit here if our sanity checks fail, but try to continue if we need to stop +if status != OCF_SUCCESS and OCF_ACTION != "stop": + exit(status) + +setup_constants() + +try: + if OCF_ACTION == "start": + status = qemu_colo_start() + elif OCF_ACTION == "stop": + status = qemu_colo_stop() + elif OCF_ACTION == "monitor": + status = qemu_colo_monitor()[0] + elif OCF_ACTION == "notify": + status = qemu_colo_notify() + elif OCF_ACTION == "promote": + status = qemu_colo_promote() + elif OCF_ACTION == "demote": + status = qemu_colo_demote() + elif OCF_ACTION == "validate-all": + status = qemu_colo_validate_all() + else: + status = OCF_ERR_UNIMPLEMENTED +except Error: + exit(OCF_ERR_GENERIC) +else: + exit(status) From patchwork Mon May 11 12:27:02 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lukas Straub X-Patchwork-Id: 11540639 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 9073C912 for ; Mon, 11 May 2020 12:42:35 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 15D55206D6 for ; Mon, 11 May 2020 12:42:35 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=web.de header.i=@web.de header.b="WmF3YA0I" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 15D55206D6 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=web.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:56644 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jY7lR-0007Jb-Rh for patchwork-qemu-devel@patchwork.kernel.org; Mon, 11 May 2020 08:42:34 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:37702) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7Wz-0006KX-Kk for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:37 -0400 Received: from mout.web.de ([212.227.17.12]:55165) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1jY7Wx-00063s-Ic for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:37 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=web.de; s=dbaedf251592; t=1589200026; bh=rnfo5wlSh1RR4WxeQMA/5WmY3QFvd9YmkSEwfqVvhyo=; h=X-UI-Sender-Class:Date:From:To:Cc:Subject:In-Reply-To:References; b=WmF3YA0IlXzJUrrf5euVybm1hB6/6xE6ZxqNwCrrjKiVeDoO4RplpDVRHOQrRHsNi gQGGzepMAQy79DDDzZ/CwqxXDwcbOM0TMW9uqH9gahu4N/2/4o6xfufUJDs7QrmwIP ZVb7XDlsxiyAeXB5MfZj6uE24VJeCvvkoGR0CDeI= X-UI-Sender-Class: c548c8c5-30a9-4db5-a2e7-cb6cb037b8f9 Received: from luklap ([89.247.255.192]) by smtp.web.de (mrweb103 [213.165.67.124]) with ESMTPSA (Nemesis) id 0MOipW-1jVNxn2e6I-0066m1; Mon, 11 May 2020 14:27:05 +0200 Date: Mon, 11 May 2020 14:27:02 +0200 From: Lukas Straub To: qemu-devel Subject: [PATCH 3/5] colo: Introduce high-level test suite Message-ID: <9ebbc340436b2b60c27bdb1fc5d47734d9854514.1589199922.git.lukasstraub2@web.de> In-Reply-To: References: MIME-Version: 1.0 X-Provags-ID: V03:K1:KNomQMa/40NiyynFHRXACmb/n6e3tmf5C1DZ0tOuTb1ab6/4Lya kY1bqJrB9syOlnoKcUqeRzXkNmwlTrHLqqP0GRP/9V1aRtMYkZbvyl4yow2Qp/WBvqhXaVN QPLel4vC3oGi78+hlVCTVfiSkAe+YZFOaftogsMmiJnk+HasrDaGnQlVOKoFEDw272GssNH Y6DwxExArDwyaSNMllD8g== X-UI-Out-Filterresults: notjunk:1;V03:K0:GhXrWHrDMfs=:xB1219hGDRtFSZ/7mDcqg1 biFSkk/5y7yFSaKV0juGqeUpQ6FQ7ZAXEVyDT0lAhZ09tLNjWyWSPz5zra56+kC/1cEtJI8WR cKy3QBr92xEDHHQLL7CjAm7xEEcMUjXBftkK0UK8160XVMDOxHvF15OPMR4y2ZKXzUuAMPYhz hPPWE0QX20a4YXt+IedI6Bua8sIZKEPrnZtwuvm9yB8lhhpVzd31aQd7SlMfWoQ9hamGc+bTZ jjBNqXH626K3Ahcw7Aw/gRSQyFcEIcovLD/A8V4yX+6Env6lTqlLdMNqXXg6VbF+JyWpSPSSL ybGxKV9ZOhckp1nx5UZyBE51VIGcCnsE6GJzIwhbg6UpIXer4T3MpTNAjuCWXuvroA5rejAIa GN6I/M+d0fnNC/K6w8PGsFY+P+reVRdYjFpZHoA5bKhGyNkjRqXg+AH5TvpKJCcrNgxuqHDUP UZMiuQ2m0A6ojssLbrEAMFd6FFZvNvnxJUOYmkYJkgDku9OE955YOu0+HKv/9dtQIjEwc3kj9 WFgfemX7IPAXeviYwbFxbzLx6IM66S3XxOkSnasPuCr0KpmuokvyCYJ/c2dcexq9rZ+6n4x/D t2ou6ejEcNrZYZqhOClXsUlU8DGX//b03DMU8/8+eUfAarjgY1ZW0AhcZh3cA7Ktj0ZBAbLMs QsJrrz2R/16L9BAuaiSlbY+a+VBT2JHCXQf9dV9SV+PRyjbMRYKWLEnmx3sWDWo5TG8Iauv1s qzX16nLrJZ6Y5FWAp6T9YhD6xvdeKTqYLRQnE6aWUMDVbdcyQE0bROTerdH+VCePahvT6sd9c on8omBIIrdsedjnIDzCNE4hJnRkdigrGxhWnzrprcXTmYrp+L4LSeoUWaKAi8UjRk66CdPWJk MjgZg0xJ1qbJudhy+MjZAxqMAxhhI1Qs5WL9sCtIwtKBtXMopjHFlZjH80Xb3oI0kOgz0INcy LbG1CJUJf8W6rhbfuacCdqCUKE066c4Og3FxkvJ2zB5kHownCK1PNSgzb9I1tWbrSWhFKvSjw ztR7yIzHrTTIDer2s9td8XTNfSgUl089lknJ5cri0m1Q/gGfFVbK8cUhA5+wgCdj6rHJ2Rjll Qt7c7hq9OwIMT/mfeVYfKoK9qnd4dIqrhhOFT8d4QJ8B+q8hN5QCJ4U2bC+wzTazOb1EnsPPT 6y+oJwhd27fx+SGyBMV70JOC857RQRhrRxxR+H63bS761mcD6pBra1vxSF+0nvXhRCBnpdDqL TliuLnoDy08ZetrOV Received-SPF: pass client-ip=212.227.17.12; envelope-from=lukasstraub2@web.de; helo=mout.web.de X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/11 07:10:54 X-ACL-Warn: Detected OS = Linux 2.2.x-3.x [generic] X-Spam_score_int: -24 X-Spam_score: -2.5 X-Spam_bar: -- X-Spam_report: (-2.5 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_ENVFROM_END_DIGIT=0.25, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_PASS=-0.001, T_FILL_THIS_FORM_SHORT=0.01, URIBL_BLOCKED=0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Zhang Chen , Alberto Garcia , "Dr. David Alan Gilbert" Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" Add high-level test relying on the colo resource-agent to test all failover cases while checking guest network connectivity. Signed-off-by: Lukas Straub --- scripts/colo-resource-agent/crm_master | 44 ++ scripts/colo-resource-agent/crm_resource | 12 + tests/acceptance/colo.py | 689 +++++++++++++++++++++++ 3 files changed, 745 insertions(+) create mode 100755 scripts/colo-resource-agent/crm_master create mode 100755 scripts/colo-resource-agent/crm_resource create mode 100644 tests/acceptance/colo.py diff --git a/scripts/colo-resource-agent/crm_master b/scripts/colo-resource-agent/crm_master new file mode 100755 index 0000000000..886f523bda --- /dev/null +++ b/scripts/colo-resource-agent/crm_master @@ -0,0 +1,44 @@ +#!/bin/bash + +# Fake crm_master for COLO testing +# +# Copyright (c) Lukas Straub +# +# This work is licensed under the terms of the GNU GPL, version 2 or +# later. See the COPYING file in the top-level directory. + +TMPDIR="$HA_RSCTMP" +score=0 +query=0 + +OPTIND=1 +while getopts 'Qql:Dv:N:G' opt; do + case "$opt" in + Q|q) + # Noop + ;; + "l") + # Noop + ;; + "D") + score=0 + ;; + "v") + score=$OPTARG + ;; + "N") + TMPDIR="$COLO_TEST_REMOTE_TMP" + ;; + "G") + query=1 + ;; + esac +done + +if (( query )); then + cat "${TMPDIR}/master_score" || exit 1 +else + echo $score > "${TMPDIR}/master_score" || exit 1 +fi + +exit 0 diff --git a/scripts/colo-resource-agent/crm_resource b/scripts/colo-resource-agent/crm_resource new file mode 100755 index 0000000000..ad69ff3c6b --- /dev/null +++ b/scripts/colo-resource-agent/crm_resource @@ -0,0 +1,12 @@ +#!/bin/sh + +# Fake crm_resource for COLO testing +# +# Copyright (c) Lukas Straub +# +# This work is licensed under the terms of the GNU GPL, version 2 or +# later. See the COPYING file in the top-level directory. + +# Noop + +exit 0 diff --git a/tests/acceptance/colo.py b/tests/acceptance/colo.py new file mode 100644 index 0000000000..465513fb6c --- /dev/null +++ b/tests/acceptance/colo.py @@ -0,0 +1,689 @@ +# High-level test suite for qemu COLO testing all failover cases while checking +# guest network connectivity +# +# Copyright (c) Lukas Straub +# +# This work is licensed under the terms of the GNU GPL, version 2 or +# later. See the COPYING file in the top-level directory. + +# HOWTO: +# +# This test has the following parameters: +# bridge_name: name of the bridge interface to connect qemu to +# host_address: ip address of the bridge interface +# guest_address: ip address that the guest gets from the dhcp server +# bridge_helper: path to the brige helper +# (default: /usr/lib/qemu/qemu-bridge-helper) +# install_cmd: command to run to install iperf3 and memtester in the guest +# (default: "sudo -n dnf -q -y install iperf3 memtester") +# +# To run the network tests, you have to specify the parameters. +# +# Example for running the colo tests: +# make check-acceptance FEDORA_31_ARCHES="x86_64" AVOCADO_TAGS="-t colo \ +# -p bridge_name=br0 -p host_address=192.168.220.1 \ +# -p guest_address=192.168.220.222" +# +# The colo tests currently only use x86_64 test vm images. With the +# FEDORA_31_ARCHES make variable as in the example, only the x86_64 images will +# be downloaded. +# +# If you're running the network tests as an unprivileged user, you need to set +# the suid bit on the bridge helper (chmod +s ). +# +# The dhcp server should assign a static ip to the guest, else the test may be +# unreliable. The Mac address for the guest is always 52:54:00:12:34:56. + + +import select +import sys +import subprocess +import shutil +import os +import signal +import os.path +import time +import tempfile + +from avocado import Test +from avocado import skipUnless +from avocado.utils import network +from avocado.utils import vmimage +from avocado.utils import cloudinit +from avocado.utils import ssh +from avocado.utils.path import find_command + +from avocado_qemu import pick_default_qemu_bin, BUILD_DIR, SOURCE_DIR +from qemu.qmp import QEMUMonitorProtocol + +def iperf3_available(): + try: + find_command("iperf3") + except CmdNotFoundError: + return False + return True + +class ColoTest(Test): + + # Constants + OCF_SUCCESS = 0 + OCF_ERR_GENERIC = 1 + OCF_ERR_ARGS = 2 + OCF_ERR_UNIMPLEMENTED = 3 + OCF_ERR_PERM = 4 + OCF_ERR_INSTALLED = 5 + OCF_ERR_CONFIGURED = 6 + OCF_NOT_RUNNING = 7 + OCF_RUNNING_MASTER = 8 + OCF_FAILED_MASTER = 9 + + HOSTA = 10 + HOSTB = 11 + + QEMU_OPTIONS = (" -display none -vga none -enable-kvm" + " -smp 2 -cpu host -m 768" + " -device e1000,mac=52:54:00:12:34:56,netdev=hn0" + " -device virtio-blk,drive=colo-disk0") + + FEDORA_VERSION = "31" + IMAGE_CHECKSUM = "e3c1b309d9203604922d6e255c2c5d098a309c2d46215d8fc026954f3c5c27a0" + IMAGE_SIZE = "4294967296b" + + hang_qemu = False + checkpoint_failover = False + traffic_procs = [] + + def get_image(self, temp_dir): + try: + return vmimage.get( + "fedora", arch="x86_64", version=self.FEDORA_VERSION, + checksum=self.IMAGE_CHECKSUM, algorithm="sha256", + cache_dir=self.cache_dirs[0], + snapshot_dir=temp_dir) + except: + self.cancel("Failed to download/prepare image") + + @skipUnless(ssh.SSH_CLIENT_BINARY, "No SSH client available") + def setUp(self): + # Qemu and qemu-img binary + default_qemu_bin = pick_default_qemu_bin() + self.QEMU_BINARY = self.params.get("qemu_bin", default=default_qemu_bin) + + # If qemu-img has been built, use it, otherwise the system wide one + # will be used. If none is available, the test will cancel. + qemu_img = os.path.join(BUILD_DIR, "qemu-img") + if not os.path.exists(qemu_img): + qemu_img = find_command("qemu-img", False) + if qemu_img is False: + self.cancel("Could not find \"qemu-img\", which is required to " + "create the bootable image") + self.QEMU_IMG_BINARY = qemu_img + vmimage.QEMU_IMG = qemu_img + + self.RESOURCE_AGENT = os.path.join(SOURCE_DIR, + "scripts/colo-resource-agent/colo") + self.ADD_PATH = os.path.join(SOURCE_DIR, "scripts/colo-resource-agent") + + # Logs + self.RA_LOG = os.path.join(self.outputdir, "resource-agent.log") + self.HOSTA_LOGDIR = os.path.join(self.outputdir, "hosta") + self.HOSTB_LOGDIR = os.path.join(self.outputdir, "hostb") + os.makedirs(self.HOSTA_LOGDIR) + os.makedirs(self.HOSTB_LOGDIR) + + # Temporary directories + # We don't use self.workdir because of unix socket path length + # limitations + self.TMPDIR = tempfile.mkdtemp() + self.TMPA = os.path.join(self.TMPDIR, "hosta") + self.TMPB = os.path.join(self.TMPDIR, "hostb") + os.makedirs(self.TMPA) + os.makedirs(self.TMPB) + + # Network + self.BRIDGE_NAME = self.params.get("bridge_name") + if self.BRIDGE_NAME: + self.HOST_ADDRESS = self.params.get("host_address") + self.GUEST_ADDRESS = self.params.get("guest_address") + self.BRIDGE_HELPER = self.params.get("bridge_helper", + default="/usr/lib/qemu/qemu-bridge-helper") + self.SSH_PORT = 22 + else: + # QEMU's hard coded usermode router address + self.HOST_ADDRESS = "10.0.2.2" + self.GUEST_ADDRESS = "127.0.0.1" + self.BRIDGE_HOSTA_PORT = network.find_free_port(address="127.0.0.1") + self.BRIDGE_HOSTB_PORT = network.find_free_port(address="127.0.0.1") + self.SSH_PORT = network.find_free_port(address="127.0.0.1") + + self.CLOUDINIT_HOME_PORT = network.find_free_port() + + # Find free port range + base_port = 1024 + while True: + base_port = network.find_free_port(start_port=base_port, + address="127.0.0.1") + if base_port == None: + self.cancel("Failed to find a free port") + for n in range(base_port, base_port +4): + if n > 65535: + break + if not network.is_port_free(n, "127.0.0.1"): + break + else: + # for loop above didn't break + break + + self.BASE_PORT = base_port + + # Disk images + self.log.info("Downloading/preparing boot image") + self.HOSTA_IMAGE = self.get_image(self.TMPA).path + self.HOSTB_IMAGE = self.get_image(self.TMPB).path + self.CLOUDINIT_ISO = os.path.join(self.TMPDIR, "cloudinit.iso") + + self.log.info("Preparing cloudinit image") + try: + cloudinit.iso(self.CLOUDINIT_ISO, self.name, + username="test", password="password", + phone_home_host=self.HOST_ADDRESS, + phone_home_port=self.CLOUDINIT_HOME_PORT) + except Exception as e: + self.cancel("Failed to prepare cloudinit image") + + self.QEMU_OPTIONS += " -cdrom %s" % self.CLOUDINIT_ISO + + # Network bridge + if not self.BRIDGE_NAME: + self.BRIDGE_PIDFILE = os.path.join(self.TMPDIR, "bridge.pid") + self.run_command(("'%s' -pidfile '%s'" + " -M none -display none -daemonize" + " -netdev user,id=host,hostfwd=tcp:127.0.0.1:%s-:22" + " -netdev socket,id=hosta,listen=127.0.0.1:%s" + " -netdev socket,id=hostb,listen=127.0.0.1:%s" + " -netdev hubport,id=hostport,hubid=0,netdev=host" + " -netdev hubport,id=porta,hubid=0,netdev=hosta" + " -netdev hubport,id=portb,hubid=0,netdev=hostb") + % (self.QEMU_BINARY, self.BRIDGE_PIDFILE, self.SSH_PORT, + self.BRIDGE_HOSTA_PORT, self.BRIDGE_HOSTB_PORT), 0) + + def tearDown(self): + try: + pid = self.read_pidfile(self.BRIDGE_PIDFILE) + if pid and self.check_pid(pid): + os.kill(pid, signal.SIGKILL) + except Exception as e: + pass + + try: + self.ra_stop(self.HOSTA) + except Exception as e: + pass + + try: + self.ra_stop(self.HOSTB) + except Exception as e: + pass + + try: + self.ssh_close() + except Exception as e: + pass + + for proc in self.traffic_procs: + try: + os.killpg(proc.pid, signal.SIGTERM) + except Exception as e: + pass + + shutil.rmtree(self.TMPDIR) + + def run_command(self, cmdline, expected_status, env=None): + proc = subprocess.Popen(cmdline, shell=True, stdout=subprocess.PIPE, + stderr=subprocess.STDOUT, + universal_newlines=True, env=env) + stdout, stderr = proc.communicate() + if proc.returncode != expected_status: + self.fail("command \"%s\" failed with code %s:\n%s" + % (cmdline, proc.returncode, stdout)) + + return proc.returncode + + def cat_line(self, path): + line="" + try: + fd = open(path, "r") + line = str.strip(fd.readline()) + fd.close() + except: + pass + return line + + def read_pidfile(self, pidfile): + try: + pid = int(self.cat_line(pidfile)) + except ValueError: + return None + else: + return pid + + def check_pid(self, pid): + try: + os.kill(pid, 0) + except OSError: + return False + else: + return True + + def ssh_open(self): + self.ssh_conn = ssh.Session(self.GUEST_ADDRESS, self.SSH_PORT, + user="test", password="password") + self.ssh_conn.DEFAULT_OPTIONS += (("UserKnownHostsFile", "/dev/null"),) + for i in range(10): + try: + if self.ssh_conn.connect(): + return + except Exception as e: + pass + time.sleep(4) + self.fail("sshd timeout") + + def ssh_ping(self): + if self.ssh_conn.cmd("echo ping").stdout != b"ping\n": + self.fail("ssh ping failed") + + def ssh_close(self): + self.ssh_conn.quit() + + def setup_base_env(self, host): + PATH = os.getenv("PATH", "") + env = { "PATH": "%s:%s" % (self.ADD_PATH, PATH), + "HA_LOGFILE": self.RA_LOG, + "OCF_RESOURCE_INSTANCE": "colo-test", + "OCF_RESKEY_CRM_meta_clone_max": "2", + "OCF_RESKEY_CRM_meta_notify": "true", + "OCF_RESKEY_CRM_meta_timeout": "30000", + "OCF_RESKEY_qemu_binary": self.QEMU_BINARY, + "OCF_RESKEY_qemu_img_binary": self.QEMU_IMG_BINARY, + "OCF_RESKEY_disk_size": str(self.IMAGE_SIZE), + "OCF_RESKEY_checkpoint_interval": "10000", + "OCF_RESKEY_base_port": str(self.BASE_PORT), + "OCF_RESKEY_debug": "2"} + + if host == self.HOSTA: + env.update({"OCF_RESKEY_options": + ("%s -qmp unix:%s,server,nowait" + " -drive if=none,id=parent0,file='%s'") + % (self.QEMU_OPTIONS, self.get_qmp_sock(host), + self.HOSTA_IMAGE), + "OCF_RESKEY_active_hidden_dir": self.TMPA, + "OCF_RESKEY_listen_address": "127.0.0.1", + "OCF_RESKEY_log_dir": self.HOSTA_LOGDIR, + "OCF_RESKEY_CRM_meta_on_node": "127.0.0.1", + "HA_RSCTMP": self.TMPA, + "COLO_TEST_REMOTE_TMP": self.TMPB}) + + else: + env.update({"OCF_RESKEY_options": + ("%s -qmp unix:%s,server,nowait" + " -drive if=none,id=parent0,file='%s'") + % (self.QEMU_OPTIONS, self.get_qmp_sock(host), + self.HOSTB_IMAGE), + "OCF_RESKEY_active_hidden_dir": self.TMPB, + "OCF_RESKEY_listen_address": "127.0.0.2", + "OCF_RESKEY_log_dir": self.HOSTB_LOGDIR, + "OCF_RESKEY_CRM_meta_on_node": "127.0.0.2", + "HA_RSCTMP": self.TMPB, + "COLO_TEST_REMOTE_TMP": self.TMPA}) + + if self.BRIDGE_NAME: + env["OCF_RESKEY_options"] += \ + " -netdev bridge,id=hn0,br=%s,helper='%s'" \ + % (self.BRIDGE_NAME, self.BRIDGE_HELPER) + else: + if host == self.HOSTA: + env["OCF_RESKEY_options"] += \ + " -netdev socket,id=hn0,connect=127.0.0.1:%s" \ + % self.BRIDGE_HOSTA_PORT + else: + env["OCF_RESKEY_options"] += \ + " -netdev socket,id=hn0,connect=127.0.0.1:%s" \ + % self.BRIDGE_HOSTB_PORT + return env + + def ra_start(self, host): + env = self.setup_base_env(host) + self.run_command(self.RESOURCE_AGENT + " start", self.OCF_SUCCESS, env) + + def ra_stop(self, host): + env = self.setup_base_env(host) + self.run_command(self.RESOURCE_AGENT + " stop", self.OCF_SUCCESS, env) + + def ra_monitor(self, host, expected_status): + env = self.setup_base_env(host) + self.run_command(self.RESOURCE_AGENT + " monitor", expected_status, env) + + def ra_promote(self, host): + env = self.setup_base_env(host) + self.run_command(self.RESOURCE_AGENT + " promote", self.OCF_SUCCESS,env) + + def ra_notify_start(self, host): + env = self.setup_base_env(host) + + env.update({"OCF_RESKEY_CRM_meta_notify_type": "post", + "OCF_RESKEY_CRM_meta_notify_operation": "start"}) + + if host == self.HOSTA: + env.update({"OCF_RESKEY_CRM_meta_notify_master_uname": "127.0.0.1", + "OCF_RESKEY_CRM_meta_notify_start_uname": "127.0.0.2"}) + else: + env.update({"OCF_RESKEY_CRM_meta_notify_master_uname": "127.0.0.2", + "OCF_RESKEY_CRM_meta_notify_start_uname": "127.0.0.1"}) + + self.run_command(self.RESOURCE_AGENT + " notify", self.OCF_SUCCESS, env) + + def ra_notify_stop(self, host): + env = self.setup_base_env(host) + + env.update({"OCF_RESKEY_CRM_meta_notify_type": "pre", + "OCF_RESKEY_CRM_meta_notify_operation": "stop"}) + + if host == self.HOSTA: + env.update({"OCF_RESKEY_CRM_meta_notify_master_uname": "127.0.0.1", + "OCF_RESKEY_CRM_meta_notify_stop_uname": "127.0.0.2"}) + else: + env.update({"OCF_RESKEY_CRM_meta_notify_master_uname": "127.0.0.2", + "OCF_RESKEY_CRM_meta_notify_stop_uname": "127.0.0.1"}) + + self.run_command(self.RESOURCE_AGENT + " notify", self.OCF_SUCCESS, env) + + def get_pid(self, host): + if host == self.HOSTA: + return self.read_pidfile(os.path.join(self.TMPA, + "colo-test-qemu.pid")) + else: + return self.read_pidfile(os.path.join(self.TMPB, + "colo-test-qemu.pid")) + + def get_master_score(self, host): + if host == self.HOSTA: + return int(self.cat_line(os.path.join(self.TMPA, "master_score"))) + else: + return int(self.cat_line(os.path.join(self.TMPB, "master_score"))) + + def get_qmp_sock(self, host): + if host == self.HOSTA: + return os.path.join(self.TMPA, "my-qmp.sock") + else: + return os.path.join(self.TMPB, "my-qmp.sock") + + + def kill_qemu_pre(self, host): + pid = self.get_pid(host) + + qmp_sock = self.get_qmp_sock(host) + + if self.checkpoint_failover: + qmp_conn = QEMUMonitorProtocol(qmp_sock) + qmp_conn.settimeout(10) + qmp_conn.connect() + while True: + event = qmp_conn.pull_event(wait=True) + if event["event"] == "STOP": + break + qmp_conn.close() + + + if pid and self.check_pid(pid): + if self.hang_qemu: + os.kill(pid, signal.SIGSTOP) + else: + os.kill(pid, signal.SIGKILL) + while self.check_pid(pid): + time.sleep(1) + + def kill_qemu_post(self, host): + pid = self.get_pid(host) + + if self.hang_qemu and pid and self.check_pid(pid): + os.kill(pid, signal.SIGKILL) + while self.check_pid(pid): + time.sleep(1) + + def prepare_guest(self): + pass + + def cycle_start(self, cycle): + pass + + def active_section(self): + return False + + def cycle_end(self, cycle): + pass + + def check_connection(self): + self.ssh_ping() + for proc in self.traffic_procs: + if proc.poll() != None: + self.fail("Traffic process died") + + def _test_colo(self, loop=1): + loop = max(loop, 1) + self.log.info("Will put logs in %s" % self.outputdir) + + self.ra_stop(self.HOSTA) + self.ra_stop(self.HOSTB) + + self.log.info("*** Startup ***") + self.ra_start(self.HOSTA) + self.ra_start(self.HOSTB) + + self.ra_monitor(self.HOSTA, self.OCF_SUCCESS) + self.ra_monitor(self.HOSTB, self.OCF_SUCCESS) + + self.log.info("*** Promoting ***") + self.ra_promote(self.HOSTA) + cloudinit.wait_for_phone_home(("0.0.0.0", self.CLOUDINIT_HOME_PORT), + self.name) + self.ssh_open() + self.prepare_guest() + + self.ra_notify_start(self.HOSTA) + + while self.get_master_score(self.HOSTB) != 100: + self.ra_monitor(self.HOSTA, self.OCF_RUNNING_MASTER) + self.ra_monitor(self.HOSTB, self.OCF_SUCCESS) + time.sleep(1) + + self.log.info("*** Replication started ***") + + self.check_connection() + + primary = self.HOSTA + secondary = self.HOSTB + + for n in range(loop): + self.cycle_start(n) + self.log.info("*** Secondary failover ***") + self.kill_qemu_pre(primary) + self.ra_notify_stop(secondary) + self.ra_monitor(secondary, self.OCF_SUCCESS) + self.ra_promote(secondary) + self.ra_monitor(secondary, self.OCF_RUNNING_MASTER) + self.kill_qemu_post(primary) + + self.check_connection() + + tmp = primary + primary = secondary + secondary = tmp + + self.log.info("*** Secondary continue replication ***") + self.ra_start(secondary) + self.ra_notify_start(primary) + + self.check_connection() + + # Wait for resync + while self.get_master_score(secondary) != 100: + self.ra_monitor(primary, self.OCF_RUNNING_MASTER) + self.ra_monitor(secondary, self.OCF_SUCCESS) + time.sleep(1) + + self.log.info("*** Replication started ***") + + self.check_connection() + + if self.active_section(): + break + + self.log.info("*** Primary failover ***") + self.kill_qemu_pre(secondary) + self.ra_monitor(primary, self.OCF_RUNNING_MASTER) + self.ra_notify_stop(primary) + self.ra_monitor(primary, self.OCF_RUNNING_MASTER) + self.kill_qemu_post(secondary) + + self.check_connection() + + self.log.info("*** Primary continue replication ***") + self.ra_start(secondary) + self.ra_notify_start(primary) + + self.check_connection() + + # Wait for resync + while self.get_master_score(secondary) != 100: + self.ra_monitor(primary, self.OCF_RUNNING_MASTER) + self.ra_monitor(secondary, self.OCF_SUCCESS) + time.sleep(1) + + self.log.info("*** Replication started ***") + + self.check_connection() + + self.cycle_end(n) + + self.ssh_close() + + self.ra_stop(self.HOSTA) + self.ra_stop(self.HOSTB) + + self.ra_monitor(self.HOSTA, self.OCF_NOT_RUNNING) + self.ra_monitor(self.HOSTB, self.OCF_NOT_RUNNING) + self.log.info("*** all ok ***") + + +class ColoQuickTest(ColoTest): + """ + :avocado: tags=colo + :avocado: tags=quick + :avocado: tags=arch:x86_64 + """ + + timeout = 300 + + def cycle_end(self, cycle): + self.log.info("Testing with peer qemu hanging" + " and failover during checkpoint") + self.hang_qemu = True + + def test_quick(self): + self.checkpoint_failover = True + self.log.info("Testing with peer qemu crashing" + " and failover during checkpoint") + self._test_colo(loop=2) + + +class ColoNetworkTest(ColoTest): + + def prepare_guest(self): + install_cmd = self.params.get("install_cmd", default= + "sudo -n dnf -q -y install iperf3 memtester") + self.ssh_conn.cmd(install_cmd) + # Use two instances to work around a bug in iperf3 + self.ssh_conn.cmd("iperf3 -sD; iperf3 -sD -p 5202") + + def _cycle_start(self, cycle): + pass + + def cycle_start(self, cycle): + self._cycle_start(cycle) + tests = [("", "client-to-server tcp"), ("-R", "server-to-client tcp")] + + self.log.info("Testing iperf %s" % tests[cycle % 2][1]) + iperf_cmd = "iperf3 %s -t 60 -i 60 --connect-timeout 30000 -c %s" \ + % (tests[cycle % 2][0], self.GUEST_ADDRESS) + proc = subprocess.Popen("while %s && %s; do sleep 1; done >>'%s' 2>&1" + % (iperf_cmd, iperf_cmd + " -p 5202", + os.path.join(self.outputdir, "iperf.log")), + shell=True, preexec_fn=os.setsid, stdin=subprocess.DEVNULL, + stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL) + self.traffic_procs.append(proc) + time.sleep(5) # Wait for iperf to get up to speed + + def cycle_end(self, cycle): + for proc in self.traffic_procs: + try: + os.killpg(proc.pid, signal.SIGTERM) + proc.wait() + except Exception as e: + pass + self.traffic_procs.clear() + time.sleep(20) + +class ColoRealNetworkTest(ColoNetworkTest): + """ + :avocado: tags=colo + :avocado: tags=slow + :avocado: tags=network_test + :avocado: tags=arch:x86_64 + """ + + timeout = 900 + + def active_section(self): + time.sleep(300) + return False + + @skipUnless(iperf3_available(), "iperf3 not available") + def test_network(self): + if not self.BRIDGE_NAME: + self.cancel("bridge options not given, will skip network test") + self.log.info("Testing with peer qemu crashing and network load") + self._test_colo(loop=2) + +class ColoStressTest(ColoNetworkTest): + """ + :avocado: tags=colo + :avocado: tags=slow + :avocado: tags=stress_test + :avocado: tags=arch:x86_64 + """ + + timeout = 1800 + + def _cycle_start(self, cycle): + if cycle == 4: + self.log.info("Stresstest with peer qemu hanging, network load" + " and failover during checkpoint") + self.checkpoint_failover = True + self.hang_qemu = True + elif cycle == 8: + self.log.info("Stresstest with peer qemu crashing and network load") + self.checkpoint_failover = False + self.hang_qemu = False + elif cycle == 12: + self.log.info("Stresstest with peer qemu hanging and network load") + self.checkpoint_failover = False + self.hang_qemu = True + + @skipUnless(iperf3_available(), "iperf3 not available") + def test_stress(self): + if not self.BRIDGE_NAME: + self.cancel("bridge options not given, will skip network test") + self.log.info("Stresstest with peer qemu crashing, network load" + " and failover during checkpoint") + self.checkpoint_failover = True + self._test_colo(loop=16) From patchwork Mon May 11 12:27:07 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lukas Straub X-Patchwork-Id: 11540613 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id A276992A for ; Mon, 11 May 2020 12:37:21 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 7922F206D6 for ; Mon, 11 May 2020 12:37:21 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=web.de header.i=@web.de header.b="q3d/oDlE" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 7922F206D6 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=web.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:52190 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jY7gO-000461-Gi for patchwork-qemu-devel@patchwork.kernel.org; Mon, 11 May 2020 08:37:20 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:37686) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7Ww-0006Gu-No for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:34 -0400 Received: from mout.web.de ([212.227.15.3]:60591) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7Wv-00063k-SG for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:34 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=web.de; s=dbaedf251592; t=1589200028; bh=Rdxixi7KV4eiUZcjalX0wrnXk1aVs2YQ0vK98vYI6WU=; h=X-UI-Sender-Class:Date:From:To:Cc:Subject:In-Reply-To:References; b=q3d/oDlEOOkPZymhC29pvlAMCIEWksXfdkjcn58o1Mlom7oNQXyB2XymNGCdk84cv 6rXInsKeWYYIad6RIDpuLmlaws7znUslqWP6rEhiPga5aLt3AZUNwg0eqxMekxB38z c2Mgjx+KToamCdHSzTL45Ph5ikIWkao6p5XRKDj4= X-UI-Sender-Class: c548c8c5-30a9-4db5-a2e7-cb6cb037b8f9 Received: from luklap ([89.247.255.192]) by smtp.web.de (mrweb003 [213.165.67.108]) with ESMTPSA (Nemesis) id 0LhNmy-1ilubk3QxM-00mf4N; Mon, 11 May 2020 14:27:08 +0200 Date: Mon, 11 May 2020 14:27:07 +0200 From: Lukas Straub To: qemu-devel Subject: [PATCH 4/5] configure,Makefile: Install colo resource-agent Message-ID: <334d8321f91f23759ef82a7b904f977003164475.1589199922.git.lukasstraub2@web.de> In-Reply-To: References: MIME-Version: 1.0 X-Provags-ID: V03:K1:kzRApTIPUtXSXuplx1eKea1e2uuM1fGZVxV6CUIhKCpSXo+GlmG ZwdaCoxJqxNfnJZB2C/fgG/xkJoDTwerSUv5RBo0dHqWbXPheMRHmmn8NI5auUyUCNZQim8 kSBvjDDTAyB5BFg3wIi2z8qu2Ps+OeXYGh0WbEMFcVNQJ3/fBGxLICtMREyP7i1AdBITAGQ D/RrsVGxey1F7H5v+5yug== X-UI-Out-Filterresults: notjunk:1;V03:K0:A88D/8D1VRs=:w2cXqlouHjy1Yz0HM9EBOV QOPXKFPmfH82uHLqR64F+5TrhGjXxDbv61nYNAqKzUaE9INnA1WN01iYmM2IWrTpyxEGnW67A lgWQNjc+Rlm/9fzU3Y4GE9zk59HpOu1y81ZoY4Pto2OhTYRksprb6IjIZC8o8Meefw5qx5gsx rEvNH4pngsFjEIZGOcLnFUV7jM7cM7JRrTlefbexy1c5S229JLBjzmptLpVJvYXEfbrTFSAlO In8hEqI1NV0G+GgItZROEDR8+xJTnItQ1/k1bn6VSaAX1ycai9rwEXcfvpiHRP8NWzCgL9Q2R BR0osjGLYVPkm/33cJ59u55xXrlBlgwUtZ2jsAd+YuncyejqOFVO2ryip1uqEhKL7GoOnCU5D QBFE/B8TZdx83SC5mRLXn9cUULckj+0z7TlRJvV7x2E7F8KBGeSDBIUNdoDaed/yqN/e40CwA fNWtwEpJ+c6wR/mkhpA9ULtl26wCOx5YnVkgOYI/PNw5p6+G2WPouqPlNfpJElTNid/FPLsF8 s5MCVHkh4EKgg/VC0ko6FxlJrLhePTRecEiHScSOxtEbfa96CI6KJsy+4j9em4Z+UVabyzWM+ xhXJ61zbKDEu3px/ONhvn9h0I/EzeG9ad6HjQ7s2xBaoLqZGOW3aZKV7/EO/+l8J5Kte4YT7g B6c+gvCTOr+ZPSQ8+h21Vs1dmwbxwPVlokOjbkkjMQcEM9g2itIhZcPm0RvAV5h1qoS8WKcus up2vlmpTDHpSR9ges2OuuxmNjXnxCy144o3dVxdNucUJ3OyEr+zyOa17DRocaunAtGYU2JV3T wwcHcO3FEHfm+dzmkyQqzre1DyEYu5LuucIRn0ZeYcvqltZHqEReZcKm4rsNTe9sPD5Qkxbp/ 2Wf/4rFduxtZnF4wOHbQQqxwtY+VCNFlGGiPLc9h0h0+a6F7kHmHhr98hnI1MdWsxXz+jnFDh Do12la1sB6DHoSmiMaHJggwDC4+PR6KtyqmMgDzvfJVM/gU2iMYZJ/xtlPKN+FVLCOBRvJvXF 4sbBkPhoGXpFGd5vqyifH3Pc5f0Xi3JUP4D1H3DyeF5LWb6XwAY6iwr8DA8KmrhQqxGZFp0TJ spliA9btU2+hki9CshlEZylATslbywwKF8067uyhBHEsEJ9rKW3OKs4Gvd7Zp0roGlJ2jawu3 gXSvrxTXE2WTG83UypElTJ0qkrN1ARPRSbeZGq3u+jP0XSXlWCrjbmOf8/p0+tDDTmQVgMTFN gRWLN85ldP6FbgLaC Received-SPF: pass client-ip=212.227.15.3; envelope-from=lukasstraub2@web.de; helo=mout.web.de X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/11 07:11:00 X-ACL-Warn: Detected OS = Linux 3.11 and newer X-Spam_score_int: -24 X-Spam_score: -2.5 X-Spam_bar: -- X-Spam_report: (-2.5 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_ENVFROM_END_DIGIT=0.25, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_PASS=-0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Zhang Chen , Alberto Garcia , "Dr. David Alan Gilbert" Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" Optionally install the resouce-agent so it gets picked up by pacemaker. Signed-off-by: Lukas Straub --- Makefile | 5 +++++ configure | 10 ++++++++++ 2 files changed, 15 insertions(+) diff --git a/Makefile b/Makefile index 8a9113e666..2ebffc4465 100644 --- a/Makefile +++ b/Makefile @@ -973,6 +973,11 @@ ifneq ($(DESCS),) $(INSTALL_DATA) "$$tmpf" \ "$(DESTDIR)$(qemu_datadir)/firmware/$$x"; \ done +endif +ifdef INSTALL_COLO_RA + mkdir -p "$(DESTDIR)$(libdir)/ocf/resource.d/qemu" + $(INSTALL_PROG) "scripts/colo-resource-agent/colo" \ + "$(DESTDIR)$(libdir)/ocf/resource.d/qemu/colo" endif for s in $(ICON_SIZES); do \ mkdir -p "$(DESTDIR)$(qemu_icondir)/hicolor/$${s}/apps"; \ diff --git a/configure b/configure index 23b5e93752..c9252030cf 100755 --- a/configure +++ b/configure @@ -430,6 +430,7 @@ softmmu="yes" linux_user="no" bsd_user="no" blobs="yes" +colo_ra="no" edk2_blobs="no" pkgversion="" pie="" @@ -1309,6 +1310,10 @@ for opt do ;; --disable-blobs) blobs="no" ;; + --disable-colo-ra) colo_ra="no" + ;; + --enable-colo-ra) colo_ra="yes" + ;; --with-pkgversion=*) pkgversion="$optarg" ;; --with-coroutine=*) coroutine="$optarg" @@ -1776,6 +1781,7 @@ Advanced options (experts only): --enable-gcov enable test coverage analysis with gcov --gcov=GCOV use specified gcov [$gcov_tool] --disable-blobs disable installing provided firmware blobs + --enable-colo-ra enable installing the COLO resource agent for pacemaker --with-vss-sdk=SDK-path enable Windows VSS support in QEMU Guest Agent --with-win-sdk=SDK-path path to Windows Platform SDK (to build VSS .tlb) --tls-priority default TLS protocol/cipher priority string @@ -6647,6 +6653,7 @@ echo "Linux AIO support $linux_aio" echo "Linux io_uring support $linux_io_uring" echo "ATTR/XATTR support $attr" echo "Install blobs $blobs" +echo "Install COLO resource agent $colo_ra" echo "KVM support $kvm" echo "HAX support $hax" echo "HVF support $hvf" @@ -7188,6 +7195,9 @@ fi if test "$blobs" = "yes" ; then echo "INSTALL_BLOBS=yes" >> $config_host_mak fi +if test "$colo_ra" = "yes" ; then + echo "INSTALL_COLO_RA=yes" >> $config_host_mak +fi if test "$iovec" = "yes" ; then echo "CONFIG_IOVEC=y" >> $config_host_mak fi From patchwork Mon May 11 12:27:11 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lukas Straub X-Patchwork-Id: 11540641 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 9BED8912 for ; Mon, 11 May 2020 12:42:37 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 72323206D6 for ; Mon, 11 May 2020 12:42:37 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=web.de header.i=@web.de header.b="cbkxlGQH" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 72323206D6 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=web.de Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:56696 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jY7lU-0007MI-EY for patchwork-qemu-devel@patchwork.kernel.org; Mon, 11 May 2020 08:42:36 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:37704) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jY7X0-0006LK-AX for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:38 -0400 Received: from mout.web.de ([212.227.17.12]:55115) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1jY7Wz-000649-9k for qemu-devel@nongnu.org; Mon, 11 May 2020 08:27:38 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=web.de; s=dbaedf251592; t=1589200033; bh=WHLl8IB544s/xbL4S9Gq3qqev/C2bzs+WF4myt75DTM=; h=X-UI-Sender-Class:Date:From:To:Cc:Subject:In-Reply-To:References; b=cbkxlGQH2T6rZ9gSIELKLDnoQK7B4CBVqK1QZJHBEQoqspsLl9S8lWvIGpWc4B6iz nrCEPk9kMfE+oTDkUxRBLZBUXsCQ/XbFJECSLYqs4KPlJIMFHDZkgemApfij10bNLM UNCN/2nxhfloXAjgQWBIR2coH56LrvYa9ubF0N+c= X-UI-Sender-Class: c548c8c5-30a9-4db5-a2e7-cb6cb037b8f9 Received: from luklap ([89.247.255.192]) by smtp.web.de (mrweb102 [213.165.67.124]) with ESMTPSA (Nemesis) id 0LeLWz-1ikXgy0UcH-00qCGE; Mon, 11 May 2020 14:27:13 +0200 Date: Mon, 11 May 2020 14:27:11 +0200 From: Lukas Straub To: qemu-devel Subject: [PATCH 5/5] MAINTAINERS: Add myself as maintainer for COLO resource agent Message-ID: <57d7784ead7a2f3be4437b14a0c69ed87bc55a91.1589199922.git.lukasstraub2@web.de> In-Reply-To: References: MIME-Version: 1.0 X-Provags-ID: V03:K1:rA8ZYAEABW72dxjlKx4b62qk8VzjlQyHSWFmMuBHAmzxmNQSYag u9zSg6qifV9j7h4+bfsIh/3gOsZjqltL/lyiBnCbdUtV7JwAfUtDcLnq++guFp2+vEFRjCM C+oZyWqv4KhBYixUVxOJtp9X1CNGdojx1EBqtevqgIAdt7qj0fnRWHQAgagV6RlBCuMP1Nf oXki2At3aT1gvJPA1fFOg== X-UI-Out-Filterresults: notjunk:1;V03:K0:G48vKEeada8=:qegn5ve0cO9B1KEGUmfjMQ RkbkkDtC5siS6EuLL6VqenUD6XrAYNnbr9hZhNODXVZh6THD2yEtHtJGPziMakXfwiGTcd5UR 7I8TVOkgX71tvD5oAWz6UnEjL9YkMXL/7QcnvgGJvgHkBpoddDQOtmKQ0jNQiOzdomB6N433m GdTH79mC5jBrdpikdwzhClcGByfr+rJSHBJRL971gILu/+K5Hiv8P5rMqmZp4DIw352Zv304O +ptC5jb2G7lIB9slCPv75Oyh7tfm2UUAThBPtEg2dp7g7xh7PB4FrBGTWPfNvSIm582q1UORp 052JM/dN8DhEa+kCdRuK93dwtxvGPwHk5Qsb0GHA4v5HCuUk77Wg5hTyX5PPVoXK9tXm7Euzt soSoq9+xHjx/vw7OAzjH2WWzEfotatAmIEo4hslr4/cDBht3z7hxG+LCMoprkcKYOiv/NdXvU lvNQoaa0UN35enh5jgo49KlGcm4hbdFshEclvDnPzrK/rGUCRlb2fKNtCniMRSR82WZrOFavb oaJyAoU4t/zgLWRDjrzpaFvL62flIy3hTvR+EOmP8xU/8SUtmjwZZaLYMw+QPwZB4x8zlDPD4 obMBjaPPeIDqMAEFdDgZOkXrYnmsE1GIqaIYbcPr197jO1dzk1WuMMaQjZPOHtpuIdDLApo09 a6Ir1sczfsCAv3DYDnk5GSF+kJiKT42ZXW8SNQt+GKz75X/Wlx0Kx6MWAjl2ifb/kvPj7A7lh QsQv4tkyizcviCXCisaJenyMOwUfJeW94SODSzNBO+kLmo8DpNa/+2r6Vl3hlsKrEsmXFAJSG +ZUwX8hUwhVFFA0xfFKOMXM1WlHrarXvZhqVhYlXdjUrmC6zhaDtjna9U7Ebx+qvfoDFNofvB fGeXNyZqUaTGlqdcW/FztQmW2QST2xPAr5KKjtQpSwFeNaRVOTJSi3lqgmKvzStjIFe/TcS+l VIw/JgQUJT5Jt1zQU0zNfd5CilmUlaeBi/mBGLni+0SjxETNFYNXkLO/tkqB58dINeA2UPKsR NeoHJXS0xBF10Epg0A5EgqzdWSwoT1zsSwOcRkgG1ozRGqZeLC5elW7flMziD+JzZ8J2xWHup i+uv+K6q1xu7SPg0crvGFHmXhiEj8xg9b9HTx5kWdd9zcsmvWprKf6Vt7XELw0PqOq5MVini8 qoi+XM5S/HU+mWMpA29y7KpuCAAMNZ4U7Kc3CiJbNrw9SHgTbtap3KLod48AiVSN4AGYjIBD+ IdzjMkE/+B2P++5L4 Received-SPF: pass client-ip=212.227.17.12; envelope-from=lukasstraub2@web.de; helo=mout.web.de X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/11 07:10:54 X-ACL-Warn: Detected OS = Linux 2.2.x-3.x [generic] X-Spam_score_int: -24 X-Spam_score: -2.5 X-Spam_bar: -- X-Spam_report: (-2.5 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_ENVFROM_END_DIGIT=0.25, FREEMAIL_FROM=0.001, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Zhang Chen , Alberto Garcia , "Dr. David Alan Gilbert" Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" While I'm not going to have much time for this, I'll still try to test and review patches. Signed-off-by: Lukas Straub --- MAINTAINERS | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 8cbc1fac2b..4c623a96e1 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2466,6 +2466,12 @@ F: net/colo* F: net/filter-rewriter.c F: net/filter-mirror.c +COLO resource agent and testing +M: Lukas Straub +S: Odd fixes +F: scripts/colo-resource-agent/* +F: tests/acceptance/colo.py + Record/replay M: Pavel Dovgalyuk R: Paolo Bonzini