Hi there,
I have a stupid problem, but i really can''t find a solution on
internet. I attempted to mount a client from the wrong IP address (the
routing table was not correct, my bad). The mount failed due to NAT, but
the client was able to contact the MDS. Now that I fixed the route I got
the following error:
Apr 27 14:37:34 asl5 kernel: Lustre:
2382:0:(client.c:1778:ptlrpc_expire_one_request()) @@@ Request
x1400402898583614 sent from MGC10.1.255.4 at tcp to NID 10.1.255.4 at tcp
has timed out for slow reply: [sent 1335530247] [real_sent
1335530247] [current 1335530254] [deadline 7s] [delay 0s]
req at ffff810016e18c00 x1400402898583614/t0(0)
o-1->MGS at MGC10.1.255.4@tcp_0:26/25 lens 240/8384 e 0 to 1 dl
1335530254 ref 2 fl Rpc:IX/ffffffff/ffffffff rc 0/-1
Apr 27 14:37:34 asl5 kernel: LustreError: 166-1: MGC10.1.255.4 at tcp:
Connection to service MGS via nid 10.1.255.4 at tcp was lost; in
progress operations using this service will fail.
Apr 27 14:37:34 asl5 kernel: LustreError: 15c-8: MGC10.1.255.4 at tcp:
The configuration from log ''lhome-client'' failed (-4).
This may be
the result of communication errors between this node and the MGS, a
bad configuration, or other errors. See the syslog for more information.
Apr 27 14:37:34 asl5 kernel: LustreError:
2394:0:(llite_lib.c:950:ll_fill_super()) Unable to process log: -4
Apr 27 14:37:34 asl5 kernel: Lustre: client ffff810016159800 umount
complete
Apr 27 14:37:34 asl5 kernel: LustreError:
2394:0:(obd_mount.c:2164:lustre_fill_super()) Unable to mount (-4)
I know a way to fix this is to use tunefs.lustre but I have to shutdown
the whole cluster to do that. I guess there must be a way to be able to
mount this client ignoring logs without the need to delete log for the
entire cluster.
This is with lustre 2.1.1
Thank you for the help. Please CC me ( tagliavini at fbk.eu ) to the answer.
Enrico
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.lustre.org/pipermail/lustre-discuss/attachments/20120427/a4a5675b/attachment.html