Periodically (like once every couple months) I'll have an ESX host (v3.0.1) become "disconnected" from Virtual Center (v2.0.1). The VMs are all still running fine on that host. However, without rebooting the server (and thus \*shutting down* every single VM on it (currently 35), there does not seem to be a way to get it to reconnect. I have:
\* Verified I can get to the host directly via the VI client
\* Verified that VC can ping it
\* Am using FQDNs when adding these to VC
\* Have tried restarting services (service mgmt-vmware restart)
\* Rebooted VC
This is becoming frustrating because even though it doesn't happen often, when it does, it seems the only way around it is to shut down the VMs and reboot the ESX host.
When VC no longer "sees" it and deems it disconnected, what else can be done to get it reconnected outside of rebooting?
You can try deleting the vpxuser as suggested in this post.
http://www.vmware.com/community/thread.jspa?messageID=733306򳁺
You can try deleting the vpxuser as suggested in this post.
http://www.vmware.com/community/thread.jspa?messageID=733306򳁺
Thanks for the reply! That thread led to some additional services to restart that took care of it w/o deleting the vpxuser.
Do you find that your ESX servers need that periodic reboot to prevent this from happening? What is your scheduled reboot cycle for ESX servers?
Try this on the ESX server:
service mgmt-vmware restart
service vmware-vpxa restart
I have some remote ESX servers (connected via VPN) that this happens to occassionally.
Those seem to be the two magic services to restart. I have them setup as an altiris job now so next time it will be a drag and drop event
ESX servers are not really like Windows Servers and don't need periodic reboots. As to why it happened it that's a tough one, could be something like a bug in the agent code or alot of different things. It seems like these two tasks need to be restarted in alot of cases of users having problems in these forums.
we have the same problem on our esx's.
i already knew about restarting mgmt-vmware, but is this really the solution??? this is a workaround, not a solution!
does really nobody have a propper solution?
greets
Have the same problem here with
vc 2.5 and esx 3.03
could be a memory starving problem from the console
look at this:
Soft/Hard limit: Edit the config.xml under /etc/vmware/hostd and add the following into the <config> section:
<hostdWarnMemInMB>200</hostdWarnMemInMB>
<hostdStopMemInMB>250</hostdStopMemInMB>
Do a restart on hostd process by:
service mgmt-vmware restart
or if you want to script it like me:
mv -f /etc/vmware/hostd/config.xml /etc/vmware/hostd/config.xml.bak
/bin/sed -e 's|</config>|\n<hostdWarnMemInMB>200</hostdWarnMemInMB>\n<hostdStopMemInMB>250</hostdStopMemInMB>\n\n</config>|g' /etc/vmware/hostd/config.xml.bak >> /etc/vmware/hostd/config.xml
service mgmt-vmware restart
iám stil testing it and wil let you now if it works...
The problem i had is that the config.xml had double value's like:
<hostdWarnMemInMB>200</hostdWarnMemInMB><hostdWarnMemInMB>200</hostdWarnMemInMB>
<hostdStopMemInMB>250</hostdStopMemInMB<hostdStopMemInMB>250</hostdStopMemInMB>
So the Hostd would not start because of it....
I do have other disconnect problems in the whole farm but that on is related to fw or vc 2.5u3 for which i wil create a new post...
http://communities.vmware.com/thread/191265
increase the memory on ESX,
Log in as root to the MUI of your ESX Server host. Para 3.0
Click on the Options tab.
Click Startup Profile.
Change the Reserved Memory parameter in the new window to 800MB.
To 3.5
ESX/configuration/memory/properties... you must REBOOT the ESX