| Message ID | 20200305122105.69184-1-zyan@redhat.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | ceph: don't request caps for idle open files | expand |
On Thu, 2020-03-05 at 20:20 +0800, Yan, Zheng wrote: > This series make cephfs client not request caps for open files that > idle for a long time. For the case that one active client and multiple > standby clients open the same file, this increase the possibility that > mds issues exclusive caps to the active client. > > Yan, Zheng (7): > ceph: always renew caps if mds_wanted is insufficient > ceph: consider inode's last read/write when calculating wanted caps > ceph: remove delay check logic from ceph_check_caps() > ceph: simplify calling of ceph_get_fmode() > ceph: update i_requested_max_size only when sending cap msg to auth mds > ceph: check all mds' caps after page writeback > ceph: calculate dir's wanted caps according to recent dirops > > fs/ceph/caps.c | 360 ++++++++++++++++------------------- > fs/ceph/dir.c | 21 +- > fs/ceph/file.c | 45 ++--- > fs/ceph/inode.c | 21 +- > fs/ceph/ioctl.c | 2 + > fs/ceph/mds_client.c | 16 +- > fs/ceph/super.h | 37 ++-- > include/linux/ceph/ceph_fs.h | 1 + > 8 files changed, 243 insertions(+), 260 deletions(-) > > changes since v2 > - make __ceph_caps_file_wanted() more readable > - add patch 5 and 6, which fix hung write during testing patch 1~4 > > changes since v3 > - don't queue delayed cap check for snap inode > - initialize ci->{last_rd,last_wr} to jiffies - 3600 * HZ > - make __ceph_caps_file_wanted() check inode type > > changes since v4 > - add patch 7, improve how to calculate dir's wanted caps > Thanks Zheng. This one seems to work just fine. Merged into ceph- client/testing branch with the following changes: - squashed patch 7 into patch 2 - cleaned up a few changelog grammatical and spelling errors - fix a small bit of whitespace damage Thanks!
This series make cephfs client not request caps for open files that idle for a long time. For the case that one active client and multiple standby clients open the same file, this increase the possibility that mds issues exclusive caps to the active client. Yan, Zheng (7): ceph: always renew caps if mds_wanted is insufficient ceph: consider inode's last read/write when calculating wanted caps ceph: remove delay check logic from ceph_check_caps() ceph: simplify calling of ceph_get_fmode() ceph: update i_requested_max_size only when sending cap msg to auth mds ceph: check all mds' caps after page writeback ceph: calculate dir's wanted caps according to recent dirops fs/ceph/caps.c | 360 ++++++++++++++++------------------- fs/ceph/dir.c | 21 +- fs/ceph/file.c | 45 ++--- fs/ceph/inode.c | 21 +- fs/ceph/ioctl.c | 2 + fs/ceph/mds_client.c | 16 +- fs/ceph/super.h | 37 ++-- include/linux/ceph/ceph_fs.h | 1 + 8 files changed, 243 insertions(+), 260 deletions(-) changes since v2 - make __ceph_caps_file_wanted() more readable - add patch 5 and 6, which fix hung write during testing patch 1~4 changes since v3 - don't queue delayed cap check for snap inode - initialize ci->{last_rd,last_wr} to jiffies - 3600 * HZ - make __ceph_caps_file_wanted() check inode type changes since v4 - add patch 7, improve how to calculate dir's wanted caps