On Thu, Mar 15, 2018 at 5:39 PM, Andrei Verovski <andreil1(a)starlett.lv> wrote:
Hi !
I have upgraded CentOS 7.4 with oVirt 4.2.1 host engine (with yum upgrade),
and its resulted in broken system - "503 Service Temporarily Unavailable"
when connecting to the host engine via web.
Service ovirt-engine failed to starts (logs attached at the bottom of this
email), other ovirt services seem to run fine.
yum update "ovirt-*-setup*” (upgrade 4.2 -> 4.2.1)
engine-setup
yum upgrade (OS upgrade)
Is this issue somehow related to JDK as described here?
https://bugzilla.redhat.com/show_bug.cgi?id=1217023
Not very likely. This is a very old bug.
New packages:
java-1.8.0-openjdk x86_64 1:1.8.0.161-0.b14.el7_4
java-1.8.0-openjdk-devel x86_64 1:1.8.0.161-0.b14.el7_4
java-1.8.0-openjdk-headless x86_64 1:1.8.0.161-0.b14.el7_4
installed packages (seem to be also 1.8x):
[root@node00 ~]# rpm -qa | grep jdk
java-1.8.0-openjdk-1.8.0.151-5.b12.el7_4.x86_64
java-1.8.0-openjdk-devel-1.8.0.151-5.b12.el7_4.x86_64
java-1.8.0-openjdk-headless-1.8.0.151-5.b12.el7_4.x86_64
copy-jdk-configs-2.2-5.el7_4.noarch
Indeed. Some versions of the engine could work with both openjdk 1.7
and 1.8, but recent versions require 1.8.
Since my host engine is actually a KVM appliance under another SuSE server,
I simply discarded it and reverted back old qcow2 image.
OK. Any chance you kept a copy, or at least more logs, for further debugging?
So this is a warning to anyone - don’t upgrade CentOS, or at least keep a
copy of disk image before ANY upgrade !
Having regular backups, both before significant changes like upgrades
and also routine, on-going, is always a good idea :-)
If you have all of server.log, we can try to understand the cause for
the failure. If not, and you try again to upgrade and if fails again,
please share all relevant logs before reverting to the backup. Thanks!
Best regards,
--
Didi