[ovirt-users] 3.6.x to 4.0.2 issues + gluster + move storage domains

Tue Aug 23 10:38:40 EDT 2016

Simone,

Thanks for the reply. I really appreciate you taking the time to answer.

Keeping the numbers the same as my original email:

1. Any idea why my cluster is showing as N/A and where can I look for the reason answer?
2. <answered inline below>
3. I changed to the vdsm user (broker user) and ran sendmail test from the command line of my host and it was delivered. Is this an adequate test?
4. <answered inline below>
5. Any ideas about: ETL service sampling error and where to look for answers?
6. Networking on my first host (the one that is running as NFS server) remains a mess. I came in on Friday *holiday here* and tried to get oVirt working again on this server. If I start vdsmd it becomes unresponsive and I can't access network on the host. As a test, I tried to setup another host with "OVS" networking instead of "Legacy" network type. It seemed fine, but would not survive a reboot... i.e. when using OVS networking, the machine would lose all access to the network after a reboot. I ended up doing a full OS wipe and reinstall on this second host. I then reinstalled oVirt with legacy networking and it is now back in my cluster. It's like OVS networking does not function on my machines. Any ideas how I can get mu first host back into the cluster (keeping in mind that it is also the NFS server for the cluster)?

Cheers,
Gervais

> On Aug 23, 2016, at 9:28 AM, Simone Tiraboschi <stirabos at redhat.com> wrote:
> 
> On Thu, Aug 18, 2016 at 7:32 PM, Gervais de Montbrun
> <gervais at demontbrun.com> wrote:
>> Hey Folks,
>> 
>> I setup an oVirt 3.6.x cluster with one of my host nodes as the NFS server.
>> I did this because I was waiting for true Gluster integration (ie oVirt 4.0)
>> and wanted to go ahead with setting up my test environment. I'd like to now
>> move my existing 3 server oVirt cluster to 4.0 with Gluster setup as for all
>> backend Storage Domains and then destroy/remove the NFS server that is
>> running on one of my nodes.
>> 
>> I've upgraded my servers to 4.0.2 with my existing setup, but I am having
>> all kinds of issues. I would like to try to clear them all up prior to
>> trying to go to Gluster for the Storage Domains and would love some help. At
>> the risk of overwhelming all of y'all, I'm going to try to list the issues I
>> am seeing below and hope that folks will let me know what they would like to
>> see in terms of logs, etc. If these are known issues that will be cleared up
>> soon (or eventually) then I will wait and apply updates are they come, but
>> want to make sure that the problems I am having are not unique to me.
>> 
>> 1. The beautiful new dashboard shows that my cluster is N/A:
>> 
>> and I see the following events:
>> 
>> I don't know which log file to search for a solution.
>> 
>> 2. My hosted engine never seems to start properly unless I manually set my
>> hosts to Global Maintenance mode. I then have to re-run "engine-setup" and
>> accept all the defaults and it will start and function. It does throw an
>> error about starting ovirt-imageio-proxy when running engine-setup
>> 
>> [ ERROR ] Failed to execute stage 'Closing up': Failed to start service
>> 'ovirt-imageio-proxy'
>> 
>> 
>> 2(a). Can I just disable ovirt-imageio-proxy somewhere? Is this just
>> supposed to make uploading ISO images simpler? I don't really need this.
> 
> It's tracked by this:
> https://bugzilla.redhat.com/show_bug.cgi?id=1365451
> 
> 4.0.3 will come with a fix
> 
>> 2(b). Do I need to have DWH enabled? I am not certain when this was added
>> into my hosted-engine setup, but I don't think I need it.
> 
> Yes, you have to since the new dashboard is consuming it.
> 
>> 3. My vdsm.log is full of entries like, "JsonRpc
>> (StompReactor)::ERROR::2016-08-18
>> 12:59:56,783::betterAsyncore::113::vds.dispatcher::(recv) SSL error during
>> reading data: unexpected eof"
> 
> Can you please check if ovirt-ha.broker can successfully reach your smtp server?
> 
>> 4. Commands throwing deprecation warning. Example:
>> 
>> [root at cultivar2 ~]# hosted-engine --vm-status
>> /usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/lib/storage_backends.py:15:
>> DeprecationWarning: vdscli uses xmlrpc. since ovirt 3.6 xmlrpc is
>> deprecated, please use vdsm.jsonrpcvdscli
>>  import vdsm.vdscli
>> 
>> or in the logs:
>> 
>> Aug 18 13:56:17 cultivar3 ovirt-ha-agent:
>> /usr/lib/python2.7/site-packages/yajsonrpc/stomp.py:352: DeprecationWarning:
>> Dispatcher.pending is deprecated. Use Dispatcher.socket.pending instead.
> 
> They are due to https://bugzilla.redhat.com/show_bug.cgi?id=1101554
> We are going to fix for 4.1
> The warning is currently harmless.
> 
>> 5. This error in the console: "ETL service sampling has encountered an
>> error. Please consult the service log for more details." ETL service
>> sampling log? Where's that?
>> 
>> 6. The host that is running the NFS server had network issues on Monday and
>> I have had to remove it from the cluster. It was working fine up until I did
>> the lastest yum updates. The networking just stopped working when vdsm
>> restarted. Could this be related to my switching the cluster "Switch type"
>> to OVS from Legacy? I have since switched it back to Legacy. I'm going to
>> completely remove oVirt and reinstall/re-add it to the cluster. As it is the
>> NFS server, I do need the network to continue to function throughout the
>> process. Any advice would be appreciated.
>> 
>> Lastly... If I clear up the above, how feasible is it to GoGO GlusterFS
>> instead of my existing straight NFS server setup that I have now? Any advice
>> on how to best move my hosted_engine, data, ISO, export Storage Domains to
>> Gluster?
>> 
>> Cheers,
>> Gervais
>> 
>> 
>> 
>> 
>> _______________________________________________
>> Users mailing list
>> Users at ovirt.org
>> http://lists.ovirt.org/mailman/listinfo/users
>> 