Hide Forgot
Created attachment 1197819 [details] engine.log Description of problem: Hello all, I have three hypervisor nodes, eachone is Ovirt node 4.0.3. On the first hypervisor I installed Ovirt hosted engine as vm following the documentation. Then from the engine I attached the second and the third hypervisor, trying to form two node cluster. The issue i'm facing is the slowness of the Ovirt Engine dashboard. Since the first login ( before adding the two hosts ) I must wait at least 10 seconds or more for any operation in the dashboard ( switching the tabs, enabling options). I tried with the latest IE and Firefox, I sometimes saw them stucking during the dashboard tasks. The first hypervisor has 24 vcpu and 32gb ram, I assigned to the engine vm 12 vcpu and 24gb ram. I also verified engine host resolution, it can resolve itself via the dns. If you want I can make some video during any task. I attach engine.log and server.log file, let me know if you need something more. Regards Alessio Dini Version-Release number of selected component (if applicable): 4.0.3 How reproducible: Just install a single Ovirt node and install inside a Hosted Engine as vm. Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Created attachment 1197820 [details] server.log
Pavel - can your team try to reproduce?
Alessio - can you test with 4.0.4?
I want to mention that ever since 4.0 introduction the DB load is way higher than it used to be. Sure, mandatory DWH has its toll, but it is much more intensive than <4.0 with reports. But is seems the higher postgres load is not only on engine_history db but also on the regular "engine" db. I don't have hard data but this should really be analyzed...
(In reply to Michal Skrivanek from comment #4) > I want to mention that ever since 4.0 introduction the DB load is way higher > than it used to be. Sure, mandatory DWH has its toll, but it is much more > intensive than <4.0 with reports. But is seems the higher postgres load is > not only on engine_history db but also on the regular "engine" db. > I don't have hard data but this should really be analyzed... The reason for that is that there was a change introduced by Shirly to reduce the polling interval of DWH to 15 seconds.... That's why it is heavier. This is configurable, as far as I know. Shirly?
(In reply to Oved Ourfali from comment #2) > Pavel - can your team try to reproduce? Also, can you check reverse resolving?
(In reply to Oved Ourfali from comment #3) > Alessio - can you test with 4.0.4? Hello, I can test with 4.0.4 in a couple of weeks, I have to remake my home-lab because in 3 weeks I will install the 4.0.4 release for a big customer. Regards
I'm closing this one for now. After re-testing please reopen if relevant.