Hi,
I am having problems with one host not logging anything past 24/03/2011 for /var/log/messages. There are still logs for var/log/vmware/vpx/vpxa.log and /var/log/vmware/hostd.log which are up-to-date.
This started happening after host was removed from cluster and replaced with newer/higher spec'd server to replace. The hostname and IP address was changed to be that of original host.
Build: ESXi 4.1 348481
- Checked logs through vSphere client and could not find any logs past 24/03/2011 (when hostname was changed)
- Host was originally introduced as a new server and not re-conected from old host
- Checked \var\log\messages folder and not logs past 24\03\2011
- Restarted management agent through DCUI
- Server was rebooted after hostname change originally but has not been rebooted since.
- Tried a services.sh restart
Do you have any ideas and preferably not rebooting server again?
Also, all other logs show time which is about 2 hours fast. All hosts connect to an NTP server and show time correctly.
Thanks,
Fadi
If you have multiple host, you may want to consider using a syslog server to host all your logs. This way, if you restart the server your logs are not flushed.
If you currently do not have a syslog server, the vMA will do it for you. The 'vilogger' command is what you will use. I currently use the vMA to log all the data from all my hosts. it's great for 10 hosts or so. Above that I would opt for something a bit more robust.
Here is an excellent write-up/article on setting up vMA to be your syslog server.
http://www.simonlong.co.uk/blog/2010/05/28/using-vma-as-your-esxi-syslog-server/
Perhaps someone was playing with the syslog options while this was still a test box. See the following KB for the locations to check:
http://kb.vmware.com/kb/1016621
Additional locations to check would be /etc/rc.local if someone was doing custom changes then running auto-backup.sh to persist across reboots. If you are the only admin then you can disregard all of the above as these don't change on their own.
mmmm....at the moment i am the only one that makes the changes in the VM space.
I have checked the knowledge base article you referred to and after following steps to configure remote syslog server with both FQDN and IP address I still cannot seem to get it working. If i navigate to the datastore path for local syslog logging i can only see logs for the 24\03\2011. Management agents have also been restarted again from the DCUI.
Do you have any other ideas? It looks like i may need to reboot.
If check the free space (vdf -h and df -h) . A restart may help, but if it doesn't you can reset the system configuration within the DCUI. After that you get a "fresh" config of ESXi and you'll need to re-register your VMs.
If you have multiple host, you may want to consider using a syslog server to host all your logs. This way, if you restart the server your logs are not flushed.
If you currently do not have a syslog server, the vMA will do it for you. The 'vilogger' command is what you will use. I currently use the vMA to log all the data from all my hosts. it's great for 10 hosts or so. Above that I would opt for something a bit more robust.
Here is an excellent write-up/article on setting up vMA to be your syslog server.
http://www.simonlong.co.uk/blog/2010/05/28/using-vma-as-your-esxi-syslog-server/