[Lustre-discuss] mdt kernel panic

[Ekaterina Popova]

Hello!
We've got MDT+2OSTS (Lustre 2.0.0). Today our MDT server suddenly broke. 
We saw - Kernel panic ...
Last record in logs are :
...
Jun 20 07:58:15 fs0001 kernel: Lustre: 
27315:0:(sec_gss.c:346:cli_ctx_expire())
ctx ffff810039e07780(0->c) get expired: 1308476346(-65949s)
Jun 20 07:58:15 fs0001 kernel: Lustre: 
27315:0:(sec.c:461:sptlrpc_req_ctx_switch
()) req ffff810038f63800: switch ctx ffff810039e07780(0->c) -> 
ffff8100a9db1980(
0->c), switch sec ffff81003656edc0(gss.keyring) -> 
ffff81003656edc0(gss.keyring)
Jun 20 07:58:15 fs0001 kernel: Lustre: 
27315:0:(sec_gss.c:1253:gss_cli_ctx_fini_
common()) reverse sec ffff81003656edc0: destroy ctx ffff810039e07780
Jun 20 08:32:24 fs0001 kernel: Lustre: 
2978:0:(gss_keyring.c:956:gss_sec_gc_ctx_
kr()) running gc
Jun 20 09:05:24 fs0001 kernel: Lustre: 
27239:0:(sec_gss.c:346:cli_ctx_expire())
ctx ffff81007ee0a0c0(0->lihep-MDT0000_UUID) get expired: 1308546323(-1s)
Jun 20 09:05:24 fs0001 kernel: LustreError: 
27239:0:(sec.c:452:sptlrpc_req_ctx_switch()) ASSERTION(req->rq_reqmsg) 
failed

Could you say, please, what's the problem and where to dig to understand 
the reason of this problem?

With regards,
Ekaterina Popova