running on rather performant
hardware including 12 cores and SSDs - but nevertheless we have load
averages >20 on it and very slow response time from the engine
IT Operations Platform Core Services
1&1 Internet SE | Brauerstraße 48
we also had a lot of issues using ILOs of older DL360 hosts. We
ended up using "ipmilan" with options "lanplus=1,power_wait=4".
Assuming of course you are able to access the management module/port
from the network your ovirt nodes are running in.
I think it depends all on what you do want to monitor.
a simple PING check is the least you should do, but this often is
You can check the management modules of your hosts for hardware
issues (iDRAC, ILO, IMM, ...) via nagios.
In Addition to that, you can install a nagios
We do have a rather large LDAP which does take some time to answer.
Using oVirt and AAA LDAP extension, I do run into timeouts.
running "ovirt-engine-extensions-tool --log-level=FINEST --log-
file=tool.log aaa login-user --profile=abc --user-name=xxx" fails
Thank you very much!
From: Ondra Machacek <omach...@redhat.com>
To: Dominique Taffin <dominique.taf...@1und1.de>, email@example.com
Subject: Re: [ovirt-users] AAA LDAP timeout
Date: Wed, 18 May 2016 12:31:56 +0200
thanks for t
is it possible to have the IP in the "host" filed have resolved from log4j so
that the hostname will be displayed there? Or have it convinced to use the
thank you and best,
I was hoping someone could give some insight details...
I have a question regarding the internal operations of ovirt-
engine. It seems the internal queue only processes up to 10 requests
/ jobs simultanious. E.g. 7 Power_ON and 3 LiveMigrations. All
remaining jobs (e.g. 50 more Power_On
We are using the log4j extension to send ovirt logs to a logstash server.
As we do have several engine hosts and only one logging backend, we do need to
filter logs by hostname. So far I am unable to provide a hostname in
Log4jLogger.properties. All Log4j configurations we have in
very nice implementation - and very good blog entry.
IMHO a very important metric should be made available: IOPS.
IOPS are a very important factor when using shared storage among a
cluster. It is a key indicator for identifying issues and performance
bottlenecks and also having a base
Mail list logo