[LU-2721] lfs fid2path logs kernel error on client when FID is not found Created: 31/Jan/13 Updated: 06/Feb/13 Resolved: 06/Feb/13 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.1.3 |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Minor |
| Reporter: | Alexander Boyko | Assignee: | WC Triage |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | patch | ||
| Rank (Obsolete): | 6615 |
| Description |
|
When probing the file system on a client with lfs fid2path command, if the FID no longer exists the following error message is dropped in kernel log: LustreError: 10725:0:(lmv_obd.c:854:lmv_iocontrol()) error: iocontrol MDC lustre-MDT0000_UUID on MDTidx 0 cmd c0086696: err = -2 Debug level appears to be set wrong for this operation. |
| Comments |
| Comment by Alexander Boyko [ 31/Jan/13 ] |
| Comment by Alexander Boyko [ 31/Jan/13 ] |
|
Xyratex MRP-673 |
| Comment by Di Wang [ 31/Jan/13 ] |
|
Note Fid2patch needs to be fixed after fid on OST and DNE are landed. http://review.whamcloud.com/#change,4345 |
| Comment by Alexander Boyko [ 01/Feb/13 ] |
|
Looks like this patch http://review.whamcloud.com/#change,4345 was updated month ago, and have problems with tests. Many errors happened if the FID no longer exist at lmv_iocontrol(). This fix change debug level for ENOENT from error to warning. |
| Comment by Di Wang [ 05/Feb/13 ] |
|
Ah, sorry, I just noticed the patch is trying to reduce the console error message here. Hmm, FID2PATH has been moved out of default section here. Could you please recheck master case OBD_IOC_FID2PATH: {
struct getinfo_fid2path *gf;
struct lmv_tgt_desc *tgt;
gf = (struct getinfo_fid2path *)karg;
tgt = lmv_find_target(lmv, &gf->gf_fid);
if (IS_ERR(tgt))
RETURN(PTR_ERR(tgt));
rc = obd_iocontrol(cmd, tgt->ltd_exp, len, karg, uarg);
break;
}
So the error message has been removed for FID2PATH. Please correct me if I am wrong here. Thanks. |
| Comment by Alexander Boyko [ 06/Feb/13 ] |
|
Patch was abandoned according to Di Wang comment. |
| Comment by Peter Jones [ 06/Feb/13 ] |
|
ok thanks Alex |