[Gluster-users] Unable to start glusterfs-server after upgrade to 3.7.6

[Pierre MGP Pro]

Hi Kaamesh,


> Hi Pierre, Im trying to downgrade to 3.6 as suggested by you and joe 
> but im having some problems:-
> I have performed the following commands
>
>
>     wget -O -
>     http://download.gluster.org/pub/gluster/glusterfs/3.6/3.6.2/Debian/wheezy/pubkey.gpg
>     | sudo apt-key add -
>     sudo sh -c "echo 'deb
>     http://download.gluster.org/pub/gluster/glusterfs/3.6/3.6.2/Debian/wheezy/apt
>     wheezy main' > /etc/apt/sources.list.d/gluster.list "
>      sudo apt-get update
>      sudo apt-get install glusterfs-server=3.6.2-2
>
>
> The commands run successfully but when i check the version of 
> glusterfs using glusterd --version, its still 3.7.6
> Any gluster command i run just prints "Segmentation Fault"
> Any ideas?
> Thanks,
> Kaamesh
OK, stop me I am wrong, but you have system based on the debian track as 
ubuntu. Here in our Department  we don't support the debian track. But, 
may be you have a sort of artefact of de- installing the preceding releases.

I get that sort of problem on my centos6.7

So my answer will be the direct answer for you. But analyse the process. 
When I want to change the glusterfs release I do the following actions :

  * first : yum list | grep gluster. then I get the list of the package
    to erase ;
  * So I erase the package with rpm -e --nodeps ; ;
  * And redo the : yum list | grep gluster, and discover that older
    package where not erase. Then I erase them with rpm -e --nodeps
    package-release

  * Then I get a proper and stable package installation ;
  * So, now I can install the new release with the following script, in
    the directory containing the good release. :
      o for i in *.rpm ;do
      o rpm -i --nodeps $i
      o ;done

Have a good attention on the /var/lib/glusterd peers files for all of 
your nodes and the /var/lib/glusterd/gluster.info file. that is  the 
only way to get back yours volume and files. May be have a specific 
attention to the release of glusterfs in all that files. For 3.6.6 its

operating-version=30600

OK You may read that that number release is the the most appropriate. 
But it is the number installed with te 3.6.6 release. There are numbers 
of answer concerning that kind of number.

I hope that these answer will help you. I will be back to my office 
tomorrow . don't hesitate to ask for more information.

Another point. If you mount your glusterfs volume on all your node don't 
use NFS. I get a lot of problems. I think that these problems arrived 
with file locking, but I have no time today to search for that. So do a 
glusterfs fuse mount.

Good luck


Pierre Léonard.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151129/cde372c7/attachment.html>