beautypg.com

10 troubleshooting slow commit messages, 10 troubleshooting slow commit messages -46 – HP StorageWorks Scalable File Share User Manual

Page 270

background image

Troubleshooting

9–46

To display the WWIDs of all HBAs in the HP SFS system, enter the following command:

# lsdbquery system ’select * from hpls_hbas’

210000e08b0eac9c 0 fc Null down1
210100e08b2eac9c 1 fc Null down1

9.27.10 Troubleshooting slow commit messages

If messages relating to slow commit operations are reported, they may be as a result of the file system being

under a particularly heavy load. However, such messages may also be an indication of an underlying

hardware issue. The following are examples of such messages:

Aug 10 09:03:31 south-adm kernel: Lustre:
4647:0:(filter_io_24.c:426:filter_commitrw_write()) south-ost3: slow commitrw
commit 15s
Aug 10 09:03:31 south-adm kernel: Lustre:
4647:0:(filter_io_24.c:426:filter_commitrw_write()) skipped 1 similar messages
(ending 24711.-412 seconds ago)
Aug 10 09:03:47 south-adm kernel: Lustre:
4633:0:(filter_io_24.c:426:filter_commitrw_write()) south-ost3: slow commitrw
commit 16s
Aug 10 09:03:47 south-adm kernel: Lustre:
4633:0:(filter_io_24.c:426:filter_commitrw_write()) skipped 1 similar messages
(ending 16.470 seconds ago)
Aug 10 09:04:03 south-adm kernel: Lustre:
4638:0:(filter_io_24.c:408:filter_commitrw_write()) south-ost3: slow brw_start 15s
Aug 10 09:04:03 south-adm kernel: Lustre:
4638:0:(filter_io_24.c:408:filter_commitrw_write()) skipped 3 similar messages
(ending 28681.-352 seconds ago)
Aug 10 09:04:03 south-adm kernel: Lustre:
4627:0:(filter_io_24.c:408:filter_commitrw_write()) south-ost3: slow brw_start 15s
Aug 10 09:04:03 south-adm kernel: Lustre:
4627:0:(filter_io_24.c:408:filter_commitrw_write()) skipped 2 similar messages
(ending 0.020 seconds ago)
Aug 10 09:04:03 south-adm kernel: Lustre:
4654:0:(filter_io_24.c:417:filter_commitrw_write()) south-ost3: slow direct_io 15s
Aug 10 09:04:03 south-adm kernel: Lustre:
4654:0:(filter_io_24.c:417:filter_commitrw_write()) skipped 14 similar messages
(ending 28676.-62 seconds ago)
Aug 10 09:04:03 south-adm kernel: Lustre:
4642:0:(filter_io_24.c:426:filter_commitrw_write()) south-ost3: slow commitrw
commit 15s
Aug 10 09:04:03 south-adm kernel: Lustre:
4642:0:(filter_io_24.c:426:filter_commitrw_write()) skipped 1 similar messages
(ending 15.940 seconds ago)
Aug 10 09:04:08 south-adm kernel: Lustre:
4627:0:(filter_io_24.c:417:filter_commitrw_write()) south-ost3: slow direct_io 20s
Aug 10 09:04:08 south-adm kernel: Lustre:
4627:0:(filter_io_24.c:417:filter_commitrw_write()) skipped 7 similar messages
(ending 5.190 seconds ago)
Aug 10 09:04:11 south-adm kernel: Lustre:
4655:0:(filter_io_24.c:426:filter_commitrw_write()) south-ost3: slow commitrw
commit 23s
Aug 10 09:04:11 south-adm kernel: Lustre:
4655:0:(filter_io_24.c:426:filter_commitrw_write()) skipped 1 similar messages
(ending 7.890 seconds ago)

If the slow commit performance is the result of a hardware issue, performance will continue to be impacted

until you resolve the issue. The following are examples of hardware issues that could result in such messages:

Disk issues; for example, a RAID rebuild may have been triggered by a failed disk.
For information about replacing disks in an SFS20 array, see Section 8.1.10. For information on the

length of time that a RAID rebuild operation takes to complete, see Appendix D.

A software RAID rebuild may be taking place on a mirrored LUN.

See Section 9.33.3 for information about RAID rebuilds on mirrored LUNs.