[LU-1017] MDS oops when running racer test

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.2.0, Lustre 2.1.2
• Affects Version/s: Lustre 2.2.0
• Severity: 3
• Assignee: Niu Yawei (Inactive)
• Reporter: James A Simmons

Description

Each time running the racer test the MDS eventually oops when running with the newest lustre code from master.

Issue Links

• Trackbacks: Changelog 2.1 Changes from version 2.1.1 to version 2.1.2 Server support for kernels: 2.6.18308.4.1.el5 (RHEL5) 2.6.32220.17.1.el6 (RHEL6) Client support for unpatched kernels: 2.6.18308.4.1.el5 (RHEL5) 2.6.32220.17.1....
• Trackbacks: Changelog 2.2 version 2.2.0 Support for networks: o2iblnd OFED 1.5.4 Server support for kernels: 2.6.32220.4.2.el6 (RHEL6) Client support for unpatched kernels: 2.6.18274.18.1.el5 (RHEL5) 2.6.32220.4.2.el6 (RHEL6) 2.6.32.360....

Comments

[Bob Glossman (Inactive)]

http://review.whamcloud.com/#change,2629
back port to b2_1

[Build Master (Inactive)]

Integrated in lustre-master » i686,client,el6,ofa #480
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = ABORTED
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Build Master (Inactive)]

Integrated in lustre-master » x86_64,client,el6,ofa #480
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = FAILURE
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Build Master (Inactive)]

Integrated in lustre-master » x86_64,server,el6,ofa #480
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = FAILURE
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Mikhail Pershin]

this patch overlaps with previous commit b9ccecd1453c5c76fe135048c39f149c241650c6 LU-1013 obdclass: lu_object_find miss to unlink object from LRU.

Fan Yong was trying to solve the same issue it seems but cover only single case and also didn't handle -EAGAIN case. This is his change:

@@ -627,12 +627,14 @@ static struct lu_object *lu_object_find_try(const struct lu_env *env,
                 bkt->lsb_busy++;
                 cfs_hash_bd_unlock(hs, &bd, 1);
                 return o;
+        } else {
+                if (!cfs_list_empty(&shadow->lo_header->loh_lru))
+                        cfs_list_del_init(&shadow->lo_header->loh_lru);
+                lprocfs_counter_incr(s->ls_stats, LU_SS_CACHE_RACE);
+                cfs_hash_bd_unlock(hs, &bd, 1);
+                lu_object_free(env, o);
+                return shadow;
         }
-
-        lprocfs_counter_incr(s->ls_stats, LU_SS_CACHE_RACE);
-        cfs_hash_bd_unlock(hs, &bd, 1);
-        lu_object_free(env, o);
-        return shadow;
 }

Now we have other code using result of htable_lookup() without checking for -EAGAIN. Meanwhile this commit is not needed at all after LU-1017 fix, because the last does list_del_init inside of htable_lookup, so we can just revert b9ccecd1453c5c76fe135048c39f149c241650c6 to solve this issue.

[Build Master (Inactive)]

Integrated in lustre-master » i686,client,el5,ofa #468
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = SUCCESS
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Build Master (Inactive)]

Integrated in lustre-master » i686,server,el5,ofa #468
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = SUCCESS
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Build Master (Inactive)]

Integrated in lustre-master » i686,server,el5,inkernel #468
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = SUCCESS
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c

[Build Master (Inactive)]

Integrated in lustre-master » i686,client,el5,inkernel #468
LU-1017 handle -EAGAIN properly in lu_object_find_try() (Revision 4ce426ab33196ff04cc7016a2dc717e53d1b4a75)

Result = SUCCESS
Oleg Drokin : 4ce426ab33196ff04cc7016a2dc717e53d1b4a75
Files :

• lustre/obdclass/lu_object.c