Re: Client un-mounting since upgrade to 3.12.9-1 version

Milind Changire <mchangir@xxxxxxxxxx> · Tue, 12 Jun 2018 19:46:56 +0530

Kashif,
Could you share the core dump via Google Drive or something similar

Also, let me know the CPU arch and OS Distribution on which you are running gluster.

If you've installed the glusterfs-debuginfo package, you'll also get the source lines in the backtrace via gdb

On Tue, Jun 12, 2018 at 5:59 PM, mohammad kashif <kashif.alig@xxxxxxxxx> wrote:
Hi Milind, Vijay 

Thanks, I have some more information now as I straced glusterd on client

138544      0.000131 mprotect(0x7f2f70785000, 4096, PROT_READ|PROT_WRITE) = 0 <0.000026>
138544      0.000128 mprotect(0x7f2f70786000, 4096, PROT_READ|PROT_WRITE) = 0 <0.000027>
138544      0.000126 mprotect(0x7f2f70787000, 4096, PROT_READ|PROT_WRITE) = 0 <0.000027>
138544      0.000124 --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_ACCERR, si_addr=0x7f2f7c60ef88} ---
138544      0.000051 --- SIGSEGV {si_signo=SIGSEGV, si_code=SI_KERNEL, si_addr=0} ---
138551      0.105048 +++ killed by SIGSEGV (core dumped) +++
138550      0.000041 +++ killed by SIGSEGV (core dumped) +++
138547      0.000008 +++ killed by SIGSEGV (core dumped) +++
138546      0.000007 +++ killed by SIGSEGV (core dumped) +++
138545      0.000007 +++ killed by SIGSEGV (core dumped) +++
138544      0.000008 +++ killed by SIGSEGV (core dumped) +++
138543      0.000007 +++ killed by SIGSEGV (core dumped) +++

As for I understand that somehow gluster is trying to access memory in appropriate manner and kernel sends SIGSEGV 

I also got the core dump. I am trying gdb first time so I am not sure whether I am using it correctly 

gdb /usr/sbin/glusterfs core.138536

It just tell me that program terminated with signal 11, segmentation fault .

The problem is not limited to one client but happening to many clients. 

I will really appreciate any help as whole file system has become unusable 

Thanks

Kashif

On Tue, Jun 12, 2018 at 12:26 PM, Milind Changire <mchangir@xxxxxxxxxx> wrote:
Kashif,
You can change the log level by:
$ gluster volume set <vol> diagnostics.brick-log-level TRACE
$ gluster volume set <vol> diagnostics.client-log-level TRACE

and see how things fare

If you want fewer logs you can change the log-level to DEBUG instead of TRACE.

On Tue, Jun 12, 2018 at 3:37 PM, mohammad kashif <kashif.alig@xxxxxxxxx> wrote:
Hi Vijay

Now it is unmounting every 30 mins ! 

The server log at /var/log/glusterfs/bricks/glusteratlas-brics001-gv0.log have this line only

2018-06-12 09:53:19.303102] I [MSGID: 115013] [server-helpers.c:289:do_fd_cleanup] 0-atlasglust-server: fd cleanup on /atlas/atlasdata/zgubic/hmumu/histograms/v14.3/Signal
[2018-06-12 09:53:19.306190] I [MSGID: 101055] [client_t.c:443:gf_client_unref] 0-atlasglust-server: Shutting down connection <server-name> -2224879-2018/06/12-09:51:01:460889-atlasglust-client-0-0-0

There is no other information. Is there any way to increase log verbosity?

on the client 

2018-06-12 09:51:01.744980] I [MSGID: 114057] [client-handshake.c:1478:select_server_supported_programs] 0-atlasglust-client-5: Using Program GlusterFS 3.3, Num (1298437), Version (330)
[2018-06-12 09:51:01.746508] I [MSGID: 114046] [client-handshake.c:1231:client_setvolume_cbk] 0-atlasglust-client-5: Connected to atlasglust-client-5, attached to remote volume '/glusteratlas/brick006/gv0'.
[2018-06-12 09:51:01.746543] I [MSGID: 114047] [client-handshake.c:1242:client_setvolume_cbk] 0-atlasglust-client-5: Server and Client lk-version numbers are not same, reopening the fds
[2018-06-12 09:51:01.746814] I [MSGID: 114035] [client-handshake.c:202:client_set_lk_version_cbk] 0-atlasglust-client-5: Server lk version = 1
[2018-06-12 09:51:01.748449] I [MSGID: 114057] [client-handshake.c:1478:select_server_supported_programs] 0-atlasglust-client-6: Using Program GlusterFS 3.3, Num (1298437), Version (330)
[2018-06-12 09:51:01.750219] I [MSGID: 114046] [client-handshake.c:1231:client_setvolume_cbk] 0-atlasglust-client-6: Connected to atlasglust-client-6, attached to remote volume '/glusteratlas/brick007/gv0'.
[2018-06-12 09:51:01.750261] I [MSGID: 114047] [client-handshake.c:1242:client_setvolume_cbk] 0-atlasglust-client-6: Server and Client lk-version numbers are not same, reopening the fds
[2018-06-12 09:51:01.750503] I [MSGID: 114035] [client-handshake.c:202:client_set_lk_version_cbk] 0-atlasglust-client-6: Server lk version = 1
[2018-06-12 09:51:01.752207] I [fuse-bridge.c:4205:fuse_init] 0-glusterfs-fuse: FUSE inited with protocol versions: glusterfs 7.24 kernel 7.14
[2018-06-12 09:51:01.752261] I [fuse-bridge.c:4835:fuse_graph_sync] 0-fuse: switched to graph 0

is there a problem with server and client 1k version?

Thanks for your help.

Kashif

On Mon, Jun 11, 2018 at 11:52 PM, Vijay Bellur <vbellur@xxxxxxxxxx> wrote:

On Mon, Jun 11, 2018 at 8:50 AM, mohammad kashif <kashif.alig@xxxxxxxxx> wrote:
Hi

Since I have updated our gluster server and client to latest version 3.12.9-1, I am having this issue of gluster getting unmounted from client very regularly. It was not a problem before update.

Its a distributed file system with no replication. We have seven servers totaling around 480TB data. Its 97% full. 

I am using following config on server

gluster volume set atlasglust features.cache-invalidation on
gluster volume set atlasglust features.cache-invalidation-timeout 600
gluster volume set atlasglust performance.stat-prefetch on
gluster volume set atlasglust performance.cache-invalidation on
gluster volume set atlasglust performance.md-cache-timeout 600
gluster volume set atlasglust performance.parallel-readdir on
gluster volume set atlasglust performance.cache-size 1GB
gluster volume set atlasglust performance.client-io-threads on
gluster volume set atlasglust cluster.lookup-optimize on
gluster volume set atlasglust performance.stat-prefetch on
gluster volume set atlasglust client.event-threads 4
gluster volume set atlasglust server.event-threads 4

clients are mounted with this option

defaults,direct-io-mode=disable,attribute-timeout=600,entry-timeout=600,negative-timeout=600,fopen-keep-cache,rw,_netdev 

I can't see anything in the log file. Can someone suggest that how to troubleshoot this issue?

Can you please share the log file? Checking for messages related to disconnections/crashes in the log file would be a good way to start troubleshooting the problem.

Thanks,
Vijay 

_______________________________________________

Gluster-users mailing list

Gluster-users@xxxxxxxxxxx

http://lists.gluster.org/mailman/listinfo/gluster-users

-- 
Milind

-- 
Milind

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://lists.gluster.org/mailman/listinfo/gluster-users