We have a cluster in our environment called xXXXX. Recently one of our virtual servers xxxxxx migrated from One ESX host
to another ESX Host.
It was then discovered that a service running on this server stopped
working properly.
The service is called the commservice and uses ports 1098 and 1099
(Port 1098 is used for LOGIN request).
The service is used to communicate to client laptops which connect to
the server to drop data off and pick data up.
The problem was that while clients could drop data off they could not
pick data up, and the app team found that running the commservice in
debug mode showed that the 'created pipes were hanging or blocked.'
After investigating the app I realised that the server had recently
been migrated from ESX04 to ESX05 as we had a situation where we had to
put ESX04 in maintenance mode. When I migrated the server back to ESX04
the commservice started to work again, so it appears that the issue is
with ESX05.
ESX04 (DL 585 G1)and ESX05 (DL 585 G2) are slightly different model
servers, but a similar server xxxxxx runs the same commservice
successfully on ESX07 which is also a DL 585 G2, so it appears the
problem is exclusive to ESX05.
Do you have any ideas what the issue could be? Please let me know the suggestion.
Greets,
I'll take a stab...
To me it sounds like less of a hardware issue and more like a firewall/port blocking issue. That could be the firewall settings on the ESX box itself or further upstream with your router and/or firewall. Does your firewall port forward to allow outside access?
Also...
You mention that ESX07 runs the app as well, but you didn't say if it has been migrated or not. It could be that your application doesn't respond to the minor interruption when the VM switches hosts. Is it possible your port connections lose their "state" long enough to cause it to fail maybe?
Did you try to restart that VM on the ESX05 host? If not I would try that to isolate the problem further. If you can migrate it there, restart the VM and it runs fine, then it sounds like the app doesn't like vMotion.
If you restart the VM and it still doesn't work, I think its a network problem.
If you can, try migrating the VM that runs on the ESX07 box to the ESX05 and see if that works. If THAT does, it could be a problem with dissimilar hardware or something just not quite right software-wise with either your VM or application.
Regards
Hugh
Thanks for your reply. I'll check and let you know the status.