[Gluster-users] Gluster HA for SMB with CTDB - IO Error

[Stewart Webb]

Hi Gluster List,

I'm trying to configure a 2 node GlusterFS replicated volume with CTDB
managing SMB failover and I'm wondering if the behaviour I am seeing is
normal...
If I am playing back a large video file from a client (both Windows and
Linux) mounting the SMB share and issue `ctdb moveip` away from the server
the client is currently reading from, the client's player reports an I/O
error.
This also occurs if I'm in the middle of copying a file to/from the
client/server.
After this failure I am still able to interact with the network share and
resume normal operations.

I have tested with the native client from a Linux host and the failover
works with after an approximately 60 second freeze.

This may well be something I need to post to the CTDB/Samba list, but I was
wondering if there is anything obvious to try.

My configuration is as follows:
(ctdb is my ctdb distributed lock volume, online is the storage volume)

[root at stor01 ctdb]# ctdb status
Number of nodes:2
pnn:0 192.168.10.161   OK (THIS NODE)
pnn:1 192.168.10.162   OK
Generation:120988770
Size:2
hash:0 lmaster:0
hash:1 lmaster:1
Recovery mode:NORMAL (0)
Recovery master:0

[root at stor01 ctdb]# ctdb ip
Public IPs on node 0
192.168.10.171 0
192.168.10.172 1
192.168.10.181 0
192.168.10.182 1

[root at stor01 ctdb]# gluster volume info

Volume Name: ctdb
Type: Replicate
Volume ID: 95d1676f-0027-4f22-93fa-97b569a8e8cb
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: stor01:/data/brick1/ctdb/b1
Brick2: stor02:/data/brick1/ctdb/b2
Options Reconfigured:
nfs.export-volumes: on
nfs.addr-namelookup: off
performance.readdir-ahead: on

Volume Name: online
Type: Replicate
Volume ID: fd345d88-881e-4b1a-834e-9635cbe6eea2
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: stor01:/data/brick1/online
Brick2: stor02:/data/brick1/online
Options Reconfigured:
nfs.rpc-auth-allow: 192.168.0.*,192.168.10.*
nfs.export-volumes: on
nfs.addr-namelookup: off
nfs.disable: off
auth.allow: 192.168.0.*,192.168.10.*
performance.readdir-ahead: on
performance.stat-prefetch: off
server.allow-insecure: on
storage.batch-fsync-delay-usec: 0


-- 
Stewart Webb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151210/7890b9de/attachment.html>