<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p><br>
</p>
<div class="moz-cite-prefix">On 06/05/19 6:43 PM, Łukasz Michalski
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:4376d725-a451-7b18-a7a1-c5285c3570b3@zork.pl">Hi,
<br>
<br>
I have problem resolving split-brain in one of my installations.
<br>
<br>
CenOS 7, glusterfs 3.10.12, replica on two nodes:
<br>
<br>
[root@ixmed1 iscsi]# gluster volume status cluster
<br>
Status of volume: cluster
<br>
Gluster process TCP Port RDMA Port
Online Pid
<br>
------------------------------------------------------------------------------
<br>
Brick ixmed2:/glusterfs-bricks/cluster/clus
<br>
ter 49153 0 Y
3028
<br>
Brick ixmed1:/glusterfs-bricks/cluster/clus
<br>
ter 49153 0 Y
2917
<br>
Self-heal Daemon on localhost N/A N/A Y
112929
<br>
Self-heal Daemon on ixmed2 N/A N/A Y
57774
<br>
<br>
Task Status of Volume cluster
<br>
------------------------------------------------------------------------------
<br>
There are no active volume tasks
<br>
<br>
When I try to access one file glusterd reports split brain:
<br>
<br>
[2019-05-06 12:36:43.785098] E [MSGID: 108008]
[afr-read-txn.c:90:afr_read_txn_refresh_done]
0-cluster-replicate-0: Failing READ on gfid
2584a0e2-c0fa-4fde-8537-5d5b6a5a4635: split-brain observed.
[Input/output error]
<br>
[2019-05-06 12:36:43.787952] E [MSGID: 108008]
[afr-read-txn.c:90:afr_read_txn_refresh_done]
0-cluster-replicate-0: Failing FGETXATTR on gfid
2584a0e2-c0fa-4fde-8537-5d5b6a5a4635: split-brain observed.
[Input/output error]
<br>
[2019-05-06 12:36:43.788778] W [MSGID: 108027]
[afr-common.c:2722:afr_discover_done] 0-cluster-replicate-0: no
read subvols for (null)
<br>
[2019-05-06 12:36:43.790123] W [fuse-bridge.c:2254:fuse_readv_cbk]
0-glusterfs-fuse: 3352501: READ => -1
gfid=2584a0e2-c0fa-4fde-8537-5d5b6a5a4635 fd=0x7fde0803f390
(Input/output error)
<br>
[2019-05-06 12:36:43.794979] W [fuse-bridge.c:2254:fuse_readv_cbk]
0-glusterfs-fuse: 3352506: READ => -1
gfid=2584a0e2-c0fa-4fde-8537-5d5b6a5a4635 fd=0x7fde08215ed0
(Input/output error)
<br>
[2019-05-06 12:36:43.800468] W [fuse-bridge.c:2254:fuse_readv_cbk]
0-glusterfs-fuse: 3352508: READ => -1
gfid=2584a0e2-c0fa-4fde-8537-5d5b6a5a4635 fd=0x7fde08215ed0
(Input/output error)
<br>
<br>
The problem is that "gluster volume heal info" hangs for 10
seconds and returns:
<br>
<br>
Not able to fetch volfile from glusterd
<br>
Volume heal failed
<br>
<br>
glfsheal.log contains:
<br>
<br>
[2019-05-06 12:40:25.589879] I [afr.c:94:fix_quorum_options]
0-cluster-replicate-0: reindeer: incoming qtype = none
<br>
[2019-05-06 12:40:25.589967] I [afr.c:116:fix_quorum_options]
0-cluster-replicate-0: reindeer: quorum_count = 0
<br>
[2019-05-06 12:40:25.593294] W [MSGID: 101174]
[graph.c:361:_log_if_unknown_option] 0-cluster-readdir-ahead:
option 'parallel-readdir' is not recognized
<br>
[2019-05-06 12:40:25.593895] I [MSGID: 104045]
[glfs-master.c:91:notify] 0-gfapi: New graph
69786d65-6431-2d32-3037-3739322d3230 (0) coming up
<br>
[2019-05-06 12:40:25.593972] I [MSGID: 114020]
[client.c:2352:notify] 0-cluster-client-0: parent translators are
ready, attempting connect on transport
<br>
[2019-05-06 12:40:25.607836] I [MSGID: 114020]
[client.c:2352:notify] 0-cluster-client-1: parent translators are
ready, attempting connect on transport
<br>
[2019-05-06 12:40:25.608556] I [rpc-clnt.c:2000:rpc_clnt_reconfig]
0-cluster-client-0: changing port to 49153 (from 0)
<br>
[2019-05-06 12:40:25.618167] I [rpc-clnt.c:2000:rpc_clnt_reconfig]
0-cluster-client-1: changing port to 49153 (from 0)
<br>
[2019-05-06 12:40:25.629595] I [MSGID: 114057]
[client-handshake.c:1451:select_server_supported_programs]
0-cluster-client-0: Using Program GlusterFS 3.3, Num (1298437),
Version (330)
<br>
[2019-05-06 12:40:25.632031] I [MSGID: 114046]
[client-handshake.c:1216:client_setvolume_cbk] 0-cluster-client-0:
Connected to cluster-client-0, attached to remote volume
'/glusterfs-bricks/cluster/cluster'.
<br>
[2019-05-06 12:40:25.632100] I [MSGID: 114047]
[client-handshake.c:1227:client_setvolume_cbk] 0-cluster-client-0:
Server and Client lk-version numbers are not same, reopening the
fds
<br>
[2019-05-06 12:40:25.632263] I [MSGID: 108005]
[afr-common.c:4817:afr_notify] 0-cluster-replicate-0: Subvolume
'cluster-client-0' came back up; going online.
<br>
[2019-05-06 12:40:25.637707] I [MSGID: 114057]
[client-handshake.c:1451:select_server_supported_programs]
0-cluster-client-1: Using Program GlusterFS 3.3, Num (1298437),
Version (330)
<br>
[2019-05-06 12:40:25.639285] I [MSGID: 114046]
[client-handshake.c:1216:client_setvolume_cbk] 0-cluster-client-1:
Connected to cluster-client-1, attached to remote volume
'/glusterfs-bricks/cluster/cluster'.
<br>
[2019-05-06 12:40:25.639341] I [MSGID: 114047]
[client-handshake.c:1227:client_setvolume_cbk] 0-cluster-client-1:
Server and Client lk-version numbers are not same, reopening the
fds
<br>
[2019-05-06 12:40:31.564407] C
[rpc-clnt-ping.c:160:rpc_clnt_ping_timer_expired]
0-cluster-client-0: server 10.0.104.26:49153 has not responded in
the last 5 seconds, disconnecting.
<br>
[2019-05-06 12:40:31.565764] C
[rpc-clnt-ping.c:160:rpc_clnt_ping_timer_expired]
0-cluster-client-1: server 10.0.7.26:49153 has not responded in
the last 5 seconds, disconnecting.
<br>
</blockquote>
<p>This seems to be a problem. Have you changed the value of
ping-timeout ? Could you share the output of `gluster volume
info`?</p>
<p>Does the same issue occur if you try to resolve the split-brain
on the gfid 2584a0e2-c0fa-4fde-8537-5d5b6a5a4635 using the <code>gluster
volume heal <VOLNAME> split-brain </code>CLI?</p>
<p>-Ravi<br>
</p>
<blockquote type="cite"
cite="mid:4376d725-a451-7b18-a7a1-c5285c3570b3@zork.pl">[2019-05-06
12:40:35.645545] I [MSGID: 114018]
[client.c:2276:client_rpc_notify] 0-cluster-client-0: disconnected
from cluster-client-0. Client process will keep trying to connect
to glusterd until brick's port is available
<br>
[2019-05-06 12:40:35.645683] I
[socket.c:3534:socket_submit_request] 0-cluster-client-0: not
connected (priv->connected = -1)
<br>
[2019-05-06 12:40:35.645755] W [rpc-clnt.c:1693:rpc_clnt_submit]
0-cluster-client-0: failed to submit rpc-request (XID: 0x7
Program: GlusterFS 3.3, ProgVers: 330, Proc: 14) to rpc-transport
(cluster-client-0)
<br>
[2019-05-06 12:40:35.645807] W [MSGID: 114031]
[client-rpc-fops.c:797:client3_3_statfs_cbk] 0-cluster-client-0:
remote operation failed [Drugi koniec nie jest połączony]
<br>
[2019-05-06 12:40:35.645887] I
[socket.c:3534:socket_submit_request] 0-cluster-client-1: not
connected (priv->connected = -1)
<br>
[2019-05-06 12:40:35.645918] W [rpc-clnt.c:1693:rpc_clnt_submit]
0-cluster-client-1: failed to submit rpc-request (XID: 0x7
Program: GlusterFS 3.3, ProgVers: 330, Proc: 14) to rpc-transport
(cluster-client-1)
<br>
[2019-05-06 12:40:35.645955] W [MSGID: 114031]
[client-rpc-fops.c:797:client3_3_statfs_cbk] 0-cluster-client-1:
remote operation failed [Drugi koniec nie jest połączony]
<br>
[2019-05-06 12:40:35.646008] W [MSGID: 109075]
[dht-diskusage.c:44:dht_du_info_cbk] 0-cluster-dht: failed to get
disk info from cluster-replicate-0 [Drugi koniec nie jest
połączony]
<br>
[2019-05-06 12:40:35.647846] I [MSGID: 114018]
[client.c:2276:client_rpc_notify] 0-cluster-client-1: disconnected
from cluster-client-1. Client process will keep trying to connect
to glusterd until brick's port is available
<br>
[2019-05-06 12:40:35.647895] E [MSGID: 108006]
[afr-common.c:4842:afr_notify] 0-cluster-replicate-0: All
subvolumes are down. Going offline until atleast one of them comes
back up.
<br>
[2019-05-06 12:40:35.647989] I [MSGID: 108006]
[afr-common.c:4984:afr_local_init] 0-cluster-replicate-0: no
subvolumes up
<br>
[2019-05-06 12:40:35.648051] I [MSGID: 108006]
[afr-common.c:4984:afr_local_init] 0-cluster-replicate-0: no
subvolumes up
<br>
[2019-05-06 12:40:35.648122] I [MSGID: 104039]
[glfs-resolve.c:902:__glfs_active_subvol] 0-cluster: first lookup
on graph 69786d65-6431-2d32-3037-3739322d3230 (0) failed (Drugi
koniec nie jest połączony) [Drugi koniec nie jest połączony]
<br>
<br>
"Drugi koniec nie jest połączony" -> Transport endpoint not
connected
<br>
<br>
On brick process side there is an connection attempt:
<br>
<br>
[2019-05-06 12:40:25.638032] I [addr.c:182:gf_auth]
0-/glusterfs-bricks/cluster/cluster: allowed = "*", received addr
= "10.0.7.26"
<br>
[2019-05-06 12:40:25.638080] I [login.c:111:gf_auth] 0-auth/login:
allowed user names: e2f4c8f4-d040-4856-b6e3-62611fbab0ea
<br>
[2019-05-06 12:40:25.638109] I [MSGID: 115029]
[server-handshake.c:695:server_setvolume] 0-cluster-server:
accepted client from
ixmed1-207792-2019/05/06-12:40:25:562982-cluster-client-1-0-0
(version: 3.10.12)
<br>
[2019-05-06 12:40:31.565931] I [MSGID: 115036]
[server.c:559:server_rpc_notify] 0-cluster-server: disconnecting
connection from
ixmed1-207792-2019/05/06-12:40:25:562982-cluster-client-1-0-0
<br>
[2019-05-06 12:40:31.566420] I [MSGID: 101055]
[client_t.c:436:gf_client_unref] 0-cluster-server: Shutting down
connection
ixmed1-207792-2019/05/06-12:40:25:562982-cluster-client-1-0-0
<br>
<br>
I am not able to use any heal command because of this problem.
<br>
<br>
I have three volumes configured on that nodes. Configuration is
identical and "gluster volume heal" command fails for all of them.
<br>
<br>
Can anyone help?
<br>
<br>
Thanks,
<br>
Łukasz
<br>
<br>
<br>
_______________________________________________
<br>
Gluster-users mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:Gluster-users@gluster.org">Gluster-users@gluster.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.gluster.org/mailman/listinfo/gluster-users">https://lists.gluster.org/mailman/listinfo/gluster-users</a><br>
</blockquote>
</body>
</html>