| Message ID | 1534844779-118784-1-git-send-email-anton.nefedov@virtuozzo.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | discard blockstats | expand |
ping do you think we might proceed with this? or is there any general doubt about the idea? thanks, On 21/8/2018 12:46 PM, Anton Nefedov wrote: > new in v4: > - patch 7: discard and write-zeroes code paths had been separated in > 34fa110e: file-posix: Fix write_zeroes with unmap on block devices. > This patch now only accounts discards that come explicitly > through .bdrv_co_pdiscard handler. > - qapi 'Since' clauses changed 3.0 -> 3.1 > > v3: http://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg03688.html > > ---- > > qmp query-blockstats provides stats info for write/read/flush ops. > > Patches 1-6 implement the similar for discard (unmap) command for scsi > and ide disks. > Discard stat "unmap_ops / unmap_bytes" is supposed to account the ops that > have completed without an error. > > However, discard operation is advisory. Specifically, > - common block layer ignores ENOTSUP error code. > That might be returned if the block driver does not support discard, > or discard has been configured to be ignored. > - format drivers such as qcow2 may ignore discard if they were configured > to ignore that, or if the corresponding area is already marked unused > (unallocated / zero clusters). > > And what is actually useful is the number of bytes actually discarded > down on the host filesystem. > To achieve that, driver-specific statistics has been added to blockstats > (patch 8). > With patch 7, file-posix driver accounts discard operations on its level too. > > query-blockstat result: > > (note the difference between blockdevice unmap and file discard stats. qcow2 > sends fewer ops down to the file as the clusters are actually unallocated > on qcow2 level) > > { > "return": [ > { > "device": "drive-scsi0-0-0-0", > "parent": { >> "discard-bytes-ok": 262144, >> "discard-nb-ok": 4, > "stats": { >> "unmap_operations": 0, >> "unmap_merged": 0, > "flush_total_time_ns": 0, > "wr_highest_offset": 8111718400, > "wr_total_time_ns": 0, > "failed_wr_operations": 0, > "failed_rd_operations": 0, > "wr_merged": 0, > "wr_bytes": 0, > "timed_stats": [ > > ], >> "failed_unmap_operations": 0, > "failed_flush_operations": 0, > "account_invalid": false, > "rd_total_time_ns": 0, >> "invalid_unmap_operations": 0, > "flush_operations": 0, > "wr_operations": 0, >> "unmap_bytes": 0, > "rd_merged": 0, > "rd_bytes": 0, >> "unmap_total_time_ns": 0, > "invalid_flush_operations": 0, > "account_failed": false, > "rd_operations": 0, > "invalid_wr_operations": 0, > "invalid_rd_operations": 0 > }, > "node-name": "#block012", >> "driver": "file", >> "discard-nb-failed": 0 > }, > "stats": { >> "unmap_operations": 860, >> "unmap_merged": 0, > "flush_total_time_ns": 21506733, > "wr_highest_offset": 13411741696, > "wr_total_time_ns": 2212749334, > "failed_wr_operations": 0, > "failed_rd_operations": 0, > "wr_merged": 0, > "wr_bytes": 3426304, > "timed_stats": [ > > ], >> "failed_unmap_operations": 0, > "failed_flush_operations": 0, > "account_invalid": true, > "rd_total_time_ns": 3617478206, >> "invalid_unmap_operations": 0, > "flush_operations": 24, > "wr_operations": 309, >> "unmap_bytes": 11949633536, > "rd_merged": 0, > "rd_bytes": 141967360, >> "unmap_total_time_ns": 14871816, > [..] > > Anton Nefedov (8): > qapi: group BlockDeviceStats fields > qapi: add unmap to BlockDeviceStats > ide: account UNMAP (TRIM) operations > scsi: store unmap offset and nb_sectors in request struct > scsi: move unmap error checking to the complete callback > scsi: account unmap operations > file-posix: account discard operations > qapi: query-blockstat: add driver specific file-posix stats > > qapi/block-core.json | 82 +++++++++++++++++++++++++++++++++++++++------- > include/block/accounting.h | 1 + > include/block/block.h | 1 + > include/block/block_int.h | 1 + > block.c | 9 +++++ > block/file-posix.c | 45 +++++++++++++++++++++++-- > block/qapi.c | 11 +++++++ > hw/ide/core.c | 12 +++++++ > hw/scsi/scsi-disk.c | 29 +++++++++------- > tests/qemu-iotests/227.out | 18 ++++++++++ > 10 files changed, 184 insertions(+), 25 deletions(-) >
ping-2 On 18/9/2018 11:12 AM, Anton Nefedov wrote: > ping > > do you think we might proceed with this? or is there any general doubt > about the idea? > > thanks, > > On 21/8/2018 12:46 PM, Anton Nefedov wrote: >> new in v4: >> - patch 7: discard and write-zeroes code paths had been separated in >> 34fa110e: file-posix: Fix write_zeroes with unmap on block >> devices. >> This patch now only accounts discards that come explicitly >> through .bdrv_co_pdiscard handler. >> - qapi 'Since' clauses changed 3.0 -> 3.1 >> >> v3: http://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg03688.html >> >> ---- >> >> qmp query-blockstats provides stats info for write/read/flush ops. >> >> Patches 1-6 implement the similar for discard (unmap) command for scsi >> and ide disks. >> Discard stat "unmap_ops / unmap_bytes" is supposed to account the ops >> that >> have completed without an error. >> >> However, discard operation is advisory. Specifically, >> - common block layer ignores ENOTSUP error code. >> That might be returned if the block driver does not support discard, >> or discard has been configured to be ignored. >> - format drivers such as qcow2 may ignore discard if they were >> configured >> to ignore that, or if the corresponding area is already marked unused >> (unallocated / zero clusters). >> >> And what is actually useful is the number of bytes actually discarded >> down on the host filesystem. >> To achieve that, driver-specific statistics has been added to blockstats >> (patch 8). >> With patch 7, file-posix driver accounts discard operations on its >> level too. >> >> query-blockstat result: >> >> (note the difference between blockdevice unmap and file discard stats. >> qcow2 >> sends fewer ops down to the file as the clusters are actually unallocated >> on qcow2 level) >> >> { >> "return": [ >> { >> "device": "drive-scsi0-0-0-0", >> "parent": { >>> "discard-bytes-ok": 262144, >>> "discard-nb-ok": 4, >> "stats": { >>> "unmap_operations": 0, >>> "unmap_merged": 0, >> "flush_total_time_ns": 0, >> "wr_highest_offset": 8111718400, >> "wr_total_time_ns": 0, >> "failed_wr_operations": 0, >> "failed_rd_operations": 0, >> "wr_merged": 0, >> "wr_bytes": 0, >> "timed_stats": [ >> ], >>> "failed_unmap_operations": 0, >> "failed_flush_operations": 0, >> "account_invalid": false, >> "rd_total_time_ns": 0, >>> "invalid_unmap_operations": 0, >> "flush_operations": 0, >> "wr_operations": 0, >>> "unmap_bytes": 0, >> "rd_merged": 0, >> "rd_bytes": 0, >>> "unmap_total_time_ns": 0, >> "invalid_flush_operations": 0, >> "account_failed": false, >> "rd_operations": 0, >> "invalid_wr_operations": 0, >> "invalid_rd_operations": 0 >> }, >> "node-name": "#block012", >>> "driver": "file", >>> "discard-nb-failed": 0 >> }, >> "stats": { >>> "unmap_operations": 860, >>> "unmap_merged": 0, >> "flush_total_time_ns": 21506733, >> "wr_highest_offset": 13411741696, >> "wr_total_time_ns": 2212749334, >> "failed_wr_operations": 0, >> "failed_rd_operations": 0, >> "wr_merged": 0, >> "wr_bytes": 3426304, >> "timed_stats": [ >> ], >>> "failed_unmap_operations": 0, >> "failed_flush_operations": 0, >> "account_invalid": true, >> "rd_total_time_ns": 3617478206, >>> "invalid_unmap_operations": 0, >> "flush_operations": 24, >> "wr_operations": 309, >>> "unmap_bytes": 11949633536, >> "rd_merged": 0, >> "rd_bytes": 141967360, >>> "unmap_total_time_ns": 14871816, >> [..] >> >> Anton Nefedov (8): >> qapi: group BlockDeviceStats fields >> qapi: add unmap to BlockDeviceStats >> ide: account UNMAP (TRIM) operations >> scsi: store unmap offset and nb_sectors in request struct >> scsi: move unmap error checking to the complete callback >> scsi: account unmap operations >> file-posix: account discard operations >> qapi: query-blockstat: add driver specific file-posix stats >> >> qapi/block-core.json | 82 >> +++++++++++++++++++++++++++++++++++++++------- >> include/block/accounting.h | 1 + >> include/block/block.h | 1 + >> include/block/block_int.h | 1 + >> block.c | 9 +++++ >> block/file-posix.c | 45 +++++++++++++++++++++++-- >> block/qapi.c | 11 +++++++ >> hw/ide/core.c | 12 +++++++ >> hw/scsi/scsi-disk.c | 29 +++++++++------- >> tests/qemu-iotests/227.out | 18 ++++++++++ >> 10 files changed, 184 insertions(+), 25 deletions(-) >>
new in v4: - patch 7: discard and write-zeroes code paths had been separated in 34fa110e: file-posix: Fix write_zeroes with unmap on block devices. This patch now only accounts discards that come explicitly through .bdrv_co_pdiscard handler. - qapi 'Since' clauses changed 3.0 -> 3.1 v3: http://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg03688.html ---- qmp query-blockstats provides stats info for write/read/flush ops. Patches 1-6 implement the similar for discard (unmap) command for scsi and ide disks. Discard stat "unmap_ops / unmap_bytes" is supposed to account the ops that have completed without an error. However, discard operation is advisory. Specifically, - common block layer ignores ENOTSUP error code. That might be returned if the block driver does not support discard, or discard has been configured to be ignored. - format drivers such as qcow2 may ignore discard if they were configured to ignore that, or if the corresponding area is already marked unused (unallocated / zero clusters). And what is actually useful is the number of bytes actually discarded down on the host filesystem. To achieve that, driver-specific statistics has been added to blockstats (patch 8). With patch 7, file-posix driver accounts discard operations on its level too. query-blockstat result: (note the difference between blockdevice unmap and file discard stats. qcow2 sends fewer ops down to the file as the clusters are actually unallocated on qcow2 level) { "return": [ { "device": "drive-scsi0-0-0-0", "parent": { > "discard-bytes-ok": 262144, > "discard-nb-ok": 4, "stats": { > "unmap_operations": 0, > "unmap_merged": 0, "flush_total_time_ns": 0, "wr_highest_offset": 8111718400, "wr_total_time_ns": 0, "failed_wr_operations": 0, "failed_rd_operations": 0, "wr_merged": 0, "wr_bytes": 0, "timed_stats": [ ], > "failed_unmap_operations": 0, "failed_flush_operations": 0, "account_invalid": false, "rd_total_time_ns": 0, > "invalid_unmap_operations": 0, "flush_operations": 0, "wr_operations": 0, > "unmap_bytes": 0, "rd_merged": 0, "rd_bytes": 0, > "unmap_total_time_ns": 0, "invalid_flush_operations": 0, "account_failed": false, "rd_operations": 0, "invalid_wr_operations": 0, "invalid_rd_operations": 0 }, "node-name": "#block012", > "driver": "file", > "discard-nb-failed": 0 }, "stats": { > "unmap_operations": 860, > "unmap_merged": 0, "flush_total_time_ns": 21506733, "wr_highest_offset": 13411741696, "wr_total_time_ns": 2212749334, "failed_wr_operations": 0, "failed_rd_operations": 0, "wr_merged": 0, "wr_bytes": 3426304, "timed_stats": [ ], > "failed_unmap_operations": 0, "failed_flush_operations": 0, "account_invalid": true, "rd_total_time_ns": 3617478206, > "invalid_unmap_operations": 0, "flush_operations": 24, "wr_operations": 309, > "unmap_bytes": 11949633536, "rd_merged": 0, "rd_bytes": 141967360, > "unmap_total_time_ns": 14871816, [..] Anton Nefedov (8): qapi: group BlockDeviceStats fields qapi: add unmap to BlockDeviceStats ide: account UNMAP (TRIM) operations scsi: store unmap offset and nb_sectors in request struct scsi: move unmap error checking to the complete callback scsi: account unmap operations file-posix: account discard operations qapi: query-blockstat: add driver specific file-posix stats qapi/block-core.json | 82 +++++++++++++++++++++++++++++++++++++++------- include/block/accounting.h | 1 + include/block/block.h | 1 + include/block/block_int.h | 1 + block.c | 9 +++++ block/file-posix.c | 45 +++++++++++++++++++++++-- block/qapi.c | 11 +++++++ hw/ide/core.c | 12 +++++++ hw/scsi/scsi-disk.c | 29 +++++++++------- tests/qemu-iotests/227.out | 18 ++++++++++ 10 files changed, 184 insertions(+), 25 deletions(-)