• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

ESXi 5 Restarts Always Causing Issues

A

Adam12176

Limp Gawd
Joined
Dec 27, 2007
Messages
205
This has been annoying me to no end, and now it's kind of getting serious.

Setup:
Two Dell R810 hosts as mentioned in previous posts. Using 4 integrated Broadcom NICs and 1 Intel ET2 Quad PCIe NIC.

Management is setup to run on vmnic0 with standby on vmnic1.
vMotion using vmnic1 with standby on vmnic 0.
VM traffic on vmnic2/3 using LACP.
iSCSI using all 4 Intel ports (vmnic 4/5/6/7).

Dell OMSA Agent for ESXi installed.

Situation:

  • Restart a host.
  • Configured DRAC for Dell vCenter Plugin.
  • When it comes back up, some services are excruciatingly slow to come back up - USB Arbitrator comes to mind as one that sat there forever.
  • After ESXi loads again, the management IP has changed to the Dell EQL Heartbeat IP. I have no idea why this happened. Absolutely no changes were made aside from assigning an IP to the DRAC.
  • Tried restarting the management network, assigning management to the correct NIC/IP, but it fails to set it. Will NOT let me set it back to the original IP of the host. Set it to another IP in the same address space.
  • End up resetting network config entirely. Able to set host back to the original IP, but vCenter will not reconnect to it. Also can't log in to host directly through client - it does show 'loading inventory' but will not progress. Host does ping at correct IP.

I can't for sure validate this sentence, but I really think since I've installed the OMSA agent, I've had a LOT more problems on reboot of these two boxes. At this point I'd say on any reboot since then the management network does not come back up on reboot - you have to restart management from DCUI. Works as soon as you do that. Next reboot - same deal.

So TL;DR - Management network is completely fubar after restarting an ESXi host. Restarting mgmt fixes it. What the eff gives?
 
First, vSphere doesn't do LACP. So...don't. That can cause issues. Second, don't put iSCSI on all 4 Intel quad-port NICs. What happens if that card dies? For your NIC teams use one onboard NIC and one add-on card NIC.

I don't do Dell servers..so question. Is the DRAC port a dedicated port or shared?
 
Dedicated Port for DRACs, not shared.

VM Traffic is using IP Hash on the two interfaces. I thought we covered that in the other thread. While I'd like to circle back on that one, the restart issues are the real issue at hand.

VPN'ed in now, problem host will ping, but no SSH. As a test I unplugged everything but the first NIC for management before I left. Same deal. Pings, won't reconnect to vCenter, no SSH.
 
What is the Dell OMSA agent supposed to do? Is it necessary? My general policy is not to install anything in the hypervisor at all so my first instinct would be to rebuild a host without the OMSA agent and see how it functions compared to the others.

If it's more stable, then rebuild the rest without the agent.
 
Agreed. It is anecdotal at this point, I can't say with conviction that it is my smoking gun, but I have a gut feeling.

At least the host was free and clear. I might just try the rebuild and see what happens tomorrow.

Oh - and it's a Dell VIB for the vCenter plugin. Seemed to work fine when it was installed, but eh. Those plugins don't seem to measure up as I try to use them. The EQL one is damn near useless.
 
Adam12176 said:
Agreed. It is anecdotal at this point, I can't say with conviction that it is my smoking gun, but I have a gut feeling.

At least the host was free and clear. I might just try the rebuild and see what happens tomorrow.

Oh - and it's a Dell VIB for the vCenter plugin. Seemed to work fine when it was installed, but eh. Those plugins don't seem to measure up as I try to use them. The EQL one is damn near useless.
Click to expand...

Yeah, then I wouldn't bother with it. The less you mess with the hypervisor, the better.
 
I'd pull that agent and see.

Maybe I missed the other thread..but you don't need LACP for IP Hashing. You just need to force the Etherchannel/Port-Channel. vSphere won't negotiate LACP.
 
Sorry. I'm fried. Etherchannel in use. This host once again throwing a fit hit my last nerve today.
 
You must log in or register to reply here.
Back
Top