On Fri, Feb 3, 2017 at 11:17 AM, Sandro Bonazzola <sbonazzo@redhat.com> wrote:


On Fri, Feb 3, 2017 at 10:54 AM, Ralf Schenk <rs@databay.de> wrote:

Hello,

I upgraded my cluster of 8 hosts with gluster storage and hosted-engine-ha. They were already Centos 7.3 and using Ovirt 4.0.6 and gluster 3.7.x packages from storage-sig testing.

I'm missing the storage listed under storage tab but this is already filed by a bug. Increasing Cluster and Storage Compability level and also "reset emulated machine" after having upgraded one host after another without the need to shutdown vm's works well. (VM's get sign that there will be changes after reboot).

Important: you also have to issue a yum update on the host for upgrading additional components like i.e. gluster to 3.8.x. I was frightened of this step but It worked well except a configuration issue I was responsible for in gluster.vol (I had "transport socket, rdma")

Bugs/Quirks so far:

1. After restarting a single VM that used RNG-Device I got an error (it was german) but like "RNG Device not supported by cluster". I hat to disable RNG Device save the settings. Again settings and enable RNG Device. Then machine boots up.
I think there is a migration step missing from /dev/random to /dev/urandom for exisiting VM's.


Tomas, Francesco, Michal, can you please follow up on this?

 

2. I'm missing any gluster specific management features as my gluster is not managable in any way from the GUI. I expected to see my gluster now in dashboard and be able to add volumes etc. What do I need to do to "import" my existing gluster (Only one volume so far) to be managable ?


Sahina, can you please follow up on this?
 

3. Three of my hosts have the hosted engine deployed for ha. First all three where marked by a crown (running was gold and others where silver). After upgrading the 3 Host deployed hosted engine ha is not active anymore.

I can't get this host back with working ovirt-ha-agent/broker. I already rebooted, manually restarted the services but It isn't able to get cluster state according to
"hosted-engine --vm-status". The other hosts state the host status as "unknown stale-data"

I already shut down all agents on all hosts and issued a "hosted-engine --reinitialize-lockspace" but that didn't help.

Agents stops working after a timeout-error according to log:

MainThread::INFO::2017-02-02 19:24:52,040::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:24:59,185::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:25:06,333::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:25:13,554::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:25:20,710::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:25:27,865::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::ERROR::2017-02-02 19:25:27,866::hosted_engine::815::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_initialize_domain_monitor) Failed to start monitoring domain (sd_uuid=7c8deaa8-be02-4aaf-b9b4-ddc8da99ad96, host_id=3): timeout during domain acquisition
MainThread::WARNING::2017-02-02 19:25:27,866::hosted_engine::469::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(start_monitoring) Error while monitoring engine: Failed to start monitoring domain (sd_uuid=7c8deaa8-be02-4aaf-b9b4-ddc8da99ad96, host_id=3): timeout during domain acquisition
MainThread::WARNING::2017-02-02 19:25:27,866::hosted_engine::472::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(start_monitoring) Unexpected error
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 443, in start_monitoring
    self._initialize_domain_monitor()
  File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 816, in _initialize_domain_monitor
    raise Exception(msg)
Exception: Failed to start monitoring domain (sd_uuid=7c8deaa8-be02-4aaf-b9b4-ddc8da99ad96, host_id=3): timeout during domain acquisition
MainThread::ERROR::2017-02-02 19:25:27,866::hosted_engine::485::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(start_monitoring) Shutting down the agent because of 3 failures in a row!
MainThread::INFO::2017-02-02 19:25:32,087::hosted_engine::841::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_get_domain_monitor_status) VDSM domain monitor status: PENDING
MainThread::INFO::2017-02-02 19:25:34,250::hosted_engine::769::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_stop_domain_monitor) Failed to stop monitoring domain (sd_uuid=7c8deaa8-be02-4aaf-b9b4-ddc8da99ad96): Storage domain is member of pool: u'domain=7c8deaa8-be02-4aaf-b9b4-ddc8da99ad96'
MainThread::INFO::2017-02-02 19:25:34,254::agent::143::ovirt_hosted_engine_ha.agent.agent.Agent::(run) Agent shutting down

Simone, Martin, can you please follow up on this?

Ralph, could you please attach vdsm logs from on of your hosts for the relevant time frame?
 


 



The gluster volume of the engine is mounted corrctly in the host and accessible. Files are also readable etc. No clue what to do.

4. Last but not least: Ovirt is still using fuse to access VM-Disks on Gluster. I know - scheduled for 4.1.1 - but it was already there in 3.5.x and was scheduled for every release since then. I had this feature with opennebula already two years ago and performance is sooo much better.... So please GET IT IN  !

We're aware of the performance increase, storage and gluster teams are working on it. Maybe Sahina or Allon may follow up with current status of the feature.


 

Bye



Am 02.02.2017 um 13:19 schrieb Sandro Bonazzola:
Hi,
did you install/update to 4.1.0? Let us know your experience!
We end up knowing only when things doesn't work well, let us know it works fine for you :-)

--


Ralf Schenk
fon +49 (0) 24 05 / 40 83 70
fax +49 (0) 24 05 / 40 83 759
mail rs@databay.de
 
Databay AG
Jens-Otto-Krag-Straße 11
D-52146 Würselen
www.databay.de

Sitz/Amtsgericht Aachen • HRB:8437 • USt-IdNr.: DE 210844202
Vorstand: Ralf Schenk, Dipl.-Ing. Jens Conze, Aresch Yavari, Dipl.-Kfm. Philipp Hermanns
Aufsichtsratsvorsitzender: Wilhelm Dohmen


_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users




--
Sandro Bonazzola
Better technology. Faster innovation. Powered by community collaboration.
See how it works at redhat.com