Gluster users - Dec 2016 - RE : Frequent connect and disconnect messages flooded in logs

Hi,

are you using the 3.7 branch since it was released or did you use 
another version before?
I don't know if it's possible in your case, but would it be feasible to 
downgrade to 3.4?
As I said, with 3.4 I did not have a single problem (well, at least no 
disconnects causing files to go into split-brain...).
It would be very interesting if the same is true in your case.

If it would help anything, I should also be able to grant access to our 
cluster.

- Micha

Am 09.12.2016 um 14:55 schrieb Riccardo Murri:
> On 2016-12-08, Micha Ober wrote:
>> There have been no reports from other  users in *this* thread until
now,
>> but I have found at least one user with a very simiar problem in an
>> older thread:
> I have posted an issue with similar symptoms on Nov. 17 under the title
> "files disappearing and re-appearing".
>
> Also GlusterFS 3.7 used in a SLURM cluster, very similar log lines,
> FUSE mounts, data servers basically idle, yet they come and go and so
> do files on the client side.
>
> I may be able to provide access to the cluster if anyone is willing to
> debug the issue.
>
> Thanks,
> Riccardo
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-users

Hello,

Micha Ober <micha2k at gmail.com> ha scritto:
> are you using the 3.7 branch since it was released or did you use 
> another version before?
The cluster was installed with 3.7, it didn't exist in 3.4-times.
(Actually, it's a short-lived cluster of VMs running on top of OpenStack.)
> I don't know if it's possible in your case, but would it be
feasible to
> downgrade to 3.4?
Not sure: (1) is 3.4 packages for Ubuntu "xenial"? 
(2) what happens to the data that is already on the volume?

You mentioned in another thread that also 3.6.9 was working fine?

I guess I'll try upgrading to 3.8.x first and then downgrade if necessary.

Thanks,
Riccardo