[lustre-discuss] Lustre errors ldlm_enqueue

[Jérôme BECOT]

Hi,

We run a lustre 2.6 server setup (centos 6) with centos 7 clients 
(2.6.92 / 2.7.0) for a while with no special trouble (except 2 ost 
crashes with a kernel panic last month).

We recently installed machines with a newer kernel, and we needed a more 
recent version of lustre client to successfully compile (2.7.65)

I'm facing a curious symptom on only one lustre client. Even with no 
activity The log is full of :

LustreError: 11-0: lustre-MDT0000-mdc-ffff8805db669000: operation 
ldlm_enqueue to node 10.0.1.60 at tcp failed: rc = -14

What does it mean ?

The other 2.7.65 machines does not complain.

I can't unload lustre modules on this machine :

[root at node61 ~]# umount /scratch/
[root at node61 ~]# lustre_rmmod
   0 UP mgc MGC10.0.1.60 at tcp c739b3da-e734-a586-6269-5fd2771cb1bc 5
   1 UP lov lustre-clilov-ffff88062c5eb800 
f5e4f5f2-cf92-be19-87ec-5194e6661463 4
   2 UP lmv lustre-clilmv-ffff88062c5eb800 
f5e4f5f2-cf92-be19-87ec-5194e6661463 4
   3 UP mdc lustre-MDT0000-mdc-ffff88062c5eb800 
f5e4f5f2-cf92-be19-87ec-5194e6661463 5
   4 UP osc lustre-OST0000-osc-ffff88062c5eb800 
f5e4f5f2-cf92-be19-87ec-5194e6661463 5
   5 UP osc lustre-OST0001-osc-ffff88062c5eb800 
f5e4f5f2-cf92-be19-87ec-5194e6661463 5
Modules still loaded:
lustre/osc/osc.o lustre/mgc/mgc.o lustre/llite/lustre.o lustre/lmv/lmv.o 
lustre/fld/fld.o lustre/mdc/mdc.o lustre/fid/fid.o lustre/lov/lov.o 
lnet/klnds/socklnd/ksocklnd.o lustre/ptlrpc/ptlrpc.o 
lustre/obdclass/obdclass.o lnet/lnet/lnet.o libcfs/libcfs/libcfs.o

Thanks

-- 
Jérome BECOT

Administrateur Systèmes et Réseaux

Molécules à visée Thérapeutique par des approches in Silico (MTi)
Univ Paris Diderot, UMRS973 Inserm
Case 013
Bât. Lamarck A, porte 412
35, rue Hélène Brion 75205 Paris Cedex 13
France

Tel : 01 57 27 83 82