ESXi 6 lockups

TeeJayHoward

Limpness Supreme
Joined
Feb 8, 2005
Messages
12,258
SOLVED! Issues with the X10SL7-F and ESXi 6.0:
=====================================
1) Do not use vmnic0 (The left NIC if looking at the motherboard from the rear faceplate). vmnic1 is A-OK. Yes, I know they're both i210ATs. I don't know what the difference is, but ESXi hates the left one on one of my two motherboards!
2) Do not use USB thumbdrives to install ESXi on. The SAS ports work.
=====================================




Every 30 seconds, ESXi 6 GA locks up on me for a full ten seconds. Any commands issued via the thick client during this time will not reach the server. If SSH'd in, any keystrokes sent will wait the full ten seconds before showing up on the screen. There are no log entries which appear to correspond to the lockup times.

Now here's the REALLY weird part. If you SSH in and run esxtop, the other SSH session will only lock up for two seconds - And those are at random.

Hardware:
X10SL7-F
16GB USB thumbdrive
ESXi 6.0 GA
16GB (4x4) Kingston ECC memory
E3-1220v3

No VMs. Not connected to a vCenter Server. This is fresh-off-the-install, right after the initial reboot.

edit: This is happening on TWO (identical except for RAM) systems!
 
Last edited:
Okay, a bit more weirdness.
SSH in. Get to a shell.

1) Type in "date". Press enter.
2) Press the up arrow. Press enter.
3) Repeat 2 once a second.

This is how I've been noting the "10 second pause every 30 seconds". I got tired of hitting up+enter, so I scripted it:

while true; do date;sleep 1;done

While this script is running, SSH doesn't freeze. Cancel the script, and it's right back to a 10 second pause every 30 seconds.

This is driving me bonkers. What's going on?


edit: correction. The issue now appears to have disappeared. WTF.

edit2: Nope, it's back. And the timers aren't synced between the two boxes. Heck, one's getting random disconnects. Gah!
 
Last edited:
Re-installing ESXi 6.0 to a 320GB WD RE2 SATA disk on the LSI2308 controller built-in to the motherboard. Let's see what that does for the issue!

edit: Well, it doesn't appear to be locking up any more. But now I'm realizing that these two identical systems with nearly identical configurations (IP address off by one) are connecting differently. ESXi-01 can't nslookup the DNS server (by name or IP). ESXi-02 can. Yet the network's obviously working, because I can SSH into either of them! They can't SSH into each other, though.

edit2: If I use vmnic1 only for the management network, EVERYTHING works great on ESXi-01. Testing it on ESXi-02.
edit3: Yup. vmnic0, if used, breaks all the stuff. If it's disconnected and unused, it all works great. This is odd, as they're both Intel i210AT NICs. Lemme try and disconnect vmnic1, see if 0 works right...
edit4: Nope. If vmnic0 is connected to anything, ESXi-01 has crappy network connectivity. If vmnic1 is connected, everything's shiny. What the hell?

edit5: Nope, not everything is working right. The thick client won't let me add ESXi-01 to the cluster.
edit6: Neither will the web client. It looks like when I switched back to the USB thumbdrive (thinking that I had found the issue), it broke the networking again!
edit7: Yup. Replaced the USB drive with the SATA drive. Now it works beautifully. Going to do the same to ESXi-02.
 
Last edited:
What is you replaced the USB drive with another one and placed it into a different USB port? Just an early AM thought.
 
What is you replaced the USB drive with another one and placed it into a different USB port? Just an early AM thought.
These boards DO have USB3 ports... I should try plugging it in to one of those instead of the 2.0 ports they were flaking out on. All of my drives are of a similar brand and size, unfortunately.
 
Back
Top