[Gluster-users] Issue with storage access when brick goes down

[L, Sridhar (Nokia - IN/Bangalore)]

Hi Bipin,

I have four hosts with replica 2. If any of the host goes down among these four, client operations will hang. And it is not exactly network ping timeout, it is ping timeout + ~20seconds.


Regards,
Sridhar L

From: EXT Bipin Kunal [mailto:bkunal at redhat.com]
Sent: Tuesday, December 08, 2015 8:45 AM
To: L, Sridhar (Nokia - IN/Bangalore) <sridhar.l at nokia.com>
Cc: gluster-users at gluster.org
Subject: Re: [Gluster-users] Issue with storage access when brick goes down


I assume that this is because the host which went down is even the host which is used for mounting client.

Suppose there are 2 host. Host1 and Host2. And client is mounted as
mount -t glusterfs host1:brick1 /mnt

In this case if host1 goes down, client will wait till network ping timeout before it starts accessing volume using other host(host2).

So I think this is expected behaviour.

Thanks,
Bipin Kunal
On Dec 7, 2015 10:57 PM, "L, Sridhar (Nokia - IN/Bangalore)" <sridhar.l at nokia.com<mailto:sridhar.l at nokia.com>> wrote:
Hello,

I am running gluster storage in distributed replicated mode. When one of the brick (host) goes offline, operations on the file system by the client will hang for a while and resumes after sometime. I searched and found that operations hang for the period set in network.ping-timeout.
Can anyone explain me why the client operations will hang even though the other brick is available with all the data?


Regards,
Sridhar L




_______________________________________________
Gluster-users mailing list
Gluster-users at gluster.org<mailto:Gluster-users at gluster.org>
http://www.gluster.org/mailman/listinfo/gluster-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151208/8bc503f1/attachment.html>