[Gluster-users] Stripe+Replicate not high available?

Christian Wittwer wittwerch at gmail.com
Tue Nov 8 07:04:59 UTC 2011


Well, it is a bug. I can write and read files, but not list all files
in a folder.

root at my-dev-os-compute4:/var/lib/nova/instances# echo "test" > foo
root at my-dev-os-compute4:/var/lib/nova/instances# cat foo
test
root at my-dev-os-compute4:/var/lib/nova/instances# ls
ls: reading directory .: Transport endpoint is not connected

Can somebody confirm that with a gluster 3.3 beta2 installation?

Regards,
Christian

2011/11/8 Christian Wittwer <wittwerch at gmail.com>:
> Hi,
> Thanks for the feedback so fare. But it still does not work as expected.
> I tried to mount the volume with this fstab:
>
> my-dev-os-compute1:/glustervmstore    /var/lib/nova/instances
> glusterfs       defaults,backupvolfile-server=my-dev-os-compute2 0 1
>
> If I mount the volume with this fstab on my-dev-os-compute4, and I'll
> do a reboot of my-dev-os-compute3, the volume is gone.
>
> root at my-dev-os-compute4:/var/lib/nova/instances# ls
> ls: reading directory .: Transport endpoint is not connected
>
> [2011-11-08 07:55:47.870661] W [fuse-bridge.c:2092:fuse_readdir_cbk]
> 0-glusterfs-fuse: 6: READDIR => -1 (Transport endpoint is not
> connected)
>
> I mean that looks like a bug? It doesn't matter which brick I
> shutdown, every time the volume seems to be gone.
>
> Regards,
> Christian
>
> 2011/11/2 M. Vale <maurovale at gmail.com>:
>>
>>
>> On 2 November 2011 03:16, Amar Tumballi <amarts at redhat.com> wrote:
>>>
>>>>
>>>> On Tue, Nov 1, 2011 at 9:26 AM, Christian Wittwer <wittwerch at gmail.com>
>>>> wrote:
>>>> > I'm mounting the volume with fstab on my-dev-os-compute2 like this:
>>>> > my-dev-os-compute1:/glustervmstore    /var/lib/nova/instances
>>>> > glusterfs       defaults        0       1
>>>>
>>>> Why you don't mount every client from localhost, if all machines join
>>>> the volume?
>>>>
>>>> localhost:/glustervmstore    /var/lib/nova/instances glusterfs defaults 0
>>>> 1
>>>>
>>>
>>> can you try option like below:
>>>
>>> my-dev-os-compute1:/glustervmstore /var/lib/nova/instances glusterfs
>>> defaults,backupvolfile-server=another-server-hostname-or-ip 0 1
>>>
>>> That should give high availability, but if both servers are down, it won't
>>> work.
>>>
>>> Regards,
>>> Amar
>>>
>>> _______________________________________________
>>> Gluster-users mailing list
>>> Gluster-users at gluster.org
>>> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>>>
>>
>>
>> Thanks it worked.
>>
>> Regards
>> MV
>>
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>>
>>
>



More information about the Gluster-users mailing list