[LU-2027] writethrough_cache_enable reports erroneous number Created: 25/Sep/12 Updated: 26/Mar/18 Resolved: 26/Mar/18 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.2.0 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Colin Faber [X] (Inactive) | Assignee: | Jian Yu |
| Resolution: | Cannot Reproduce | Votes: | 0 |
| Labels: | None | ||
| Severity: | 3 |
| Rank (Obsolete): | 4154 |
| Description |
|
When enabled (by default) /proc/fs/lustre/obdfilter/*/writethrough_cache_enable reports an erroneous number which appears to continue to increment upward. Speaking with Oleg about this, he believes it to be a bug. Per the manual this should either be 1 or 0 indicating enable or disable |
| Comments |
| Comment by Peter Jones [ 27/Sep/12 ] |
|
Yujian Could you please look into this one? Thanks Peter |
| Comment by Jian Yu [ 29/Sep/12 ] |
|
Hello Colin, I provisioned a test cluster (2 clients, 1 combined MGS/MDS and 1 OSS with 7 OSTs) with Lustre b2_2 build http://build.whamcloud.com/job/lustre-b2_2/17/ and set up a fresh Lustre filesystem by running llmount.sh. After the filesystem was up, I ran the following commands on the OSS node: [root@fat-intel-4 ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/sda1 20642428 1985276 17608576 11% / tmpfs 12298420 0 12298420 0% /dev/shm /dev/sdc5 2064208 107448 1851904 6% /mnt/ost1 /dev/sdc6 2064208 107448 1851904 6% /mnt/ost2 /dev/sdc7 2064208 107448 1851904 6% /mnt/ost3 /dev/sdc8 2064208 107448 1851904 6% /mnt/ost4 /dev/sdc9 2064208 107448 1851904 6% /mnt/ost5 /dev/sdc10 2064208 107448 1851904 6% /mnt/ost6 /dev/sdc11 2064208 107448 1851904 6% /mnt/ost7 [root@fat-intel-4 ~]# lctl get_param obdfilter.*.writethrough_cache_enable obdfilter.lustre-OST0000.writethrough_cache_enable=1 obdfilter.lustre-OST0001.writethrough_cache_enable=1 obdfilter.lustre-OST0002.writethrough_cache_enable=1 obdfilter.lustre-OST0003.writethrough_cache_enable=1 obdfilter.lustre-OST0004.writethrough_cache_enable=1 obdfilter.lustre-OST0005.writethrough_cache_enable=1 obdfilter.lustre-OST0006.writethrough_cache_enable=1 [root@fat-intel-4 ~]# lctl set_param obdfilter.*.writethrough_cache_enable 0 obdfilter.lustre-OST0000.writethrough_cache_enable=0 obdfilter.lustre-OST0001.writethrough_cache_enable=0 obdfilter.lustre-OST0002.writethrough_cache_enable=0 obdfilter.lustre-OST0003.writethrough_cache_enable=0 obdfilter.lustre-OST0004.writethrough_cache_enable=0 obdfilter.lustre-OST0005.writethrough_cache_enable=0 obdfilter.lustre-OST0006.writethrough_cache_enable=0 [root@fat-intel-4 ~]# lctl get_param obdfilter.*.writethrough_cache_enable obdfilter.lustre-OST0000.writethrough_cache_enable=0 obdfilter.lustre-OST0001.writethrough_cache_enable=0 obdfilter.lustre-OST0002.writethrough_cache_enable=0 obdfilter.lustre-OST0003.writethrough_cache_enable=0 obdfilter.lustre-OST0004.writethrough_cache_enable=0 obdfilter.lustre-OST0005.writethrough_cache_enable=0 obdfilter.lustre-OST0006.writethrough_cache_enable=0 [root@fat-intel-4 ~]# lctl set_param obdfilter.*.writethrough_cache_enable 1 obdfilter.lustre-OST0000.writethrough_cache_enable=1 obdfilter.lustre-OST0001.writethrough_cache_enable=1 obdfilter.lustre-OST0002.writethrough_cache_enable=1 obdfilter.lustre-OST0003.writethrough_cache_enable=1 obdfilter.lustre-OST0004.writethrough_cache_enable=1 obdfilter.lustre-OST0005.writethrough_cache_enable=1 obdfilter.lustre-OST0006.writethrough_cache_enable=1 [root@fat-intel-4 ~]# lctl get_param obdfilter.*.writethrough_cache_enable obdfilter.lustre-OST0000.writethrough_cache_enable=1 obdfilter.lustre-OST0001.writethrough_cache_enable=1 obdfilter.lustre-OST0002.writethrough_cache_enable=1 obdfilter.lustre-OST0003.writethrough_cache_enable=1 obdfilter.lustre-OST0004.writethrough_cache_enable=1 obdfilter.lustre-OST0005.writethrough_cache_enable=1 obdfilter.lustre-OST0006.writethrough_cache_enable=1 As we can see, the value of writethrough_cache_enable was correct. I also did the same test on the following Lustre builds: All worked properly. Could you please tell me what tests/operations you performed to hit the erroneous number issue? Thanks. |
| Comment by Peter Jones [ 29/Sep/12 ] |
|
Colin Could you also please clarify whether this is a vanilla 2.2 release or whether there have been any modifications made to it? Thanks Peter |
| Comment by Jian Yu [ 26/Mar/18 ] |
|
Close old ticket. If the issue occurs again, please feel free to reopen the ticket. |