Web interface keeps dying.

[Joe Mays]

I have a really weird problem for which I have no explanation. Last week I changed the ethernet switch that a freenas box was using from a 100mbps switch to a gigabit switch, thus allowing the gigabit port on the freenas box to use its gigabit capability. Nothing else was changed, and nothing was changed on the freenas server at all.

Since then, web interface on the freenas server keeps dying. I can restart the server, and the console menu says the web ui is availalable on the correct address and port, and the web UI works immediately after boot. Sometime after that, withing 10 minutes or so, the web interface stops responding, and if you hit return on the console to reload the menu it pauses for a long time before reloading and then sends a message saying it's unable to access the web interface. Ssh access to the box also fails.

The box continues to serve nfs fine, though. It does its job, you just can't pull up the interface.

More puzzling still, we have a duplicate machine that acts as a backup server. It's the same in every way, and I also moved it to the new switch, and it is not having this problem.

Both systems are running 11.1-U6.

 

[dlavigne]

Anything in /var/log/messages around the time of the failure?

 

[Chris Moore]

I think the change to a new network switch is coincidental to the failure you are having. This sounds like a failing boot drive. Please provide a hardware rundown per this guidance with particular emphasis on hardware details around the boot drive:
https://forums.freenas.org/index.php?threads/updated-forum-rules-8-3-16.45124/

 

[Joe Mays]

I think the change to a new network switch is coincidental to the failure you are having. This sounds like a failing boot drive. Please provide a hardware rundown per this guidance with particular emphasis on hardware details around the boot drive:
https://forums.freenas.org/index.php?threads/updated-forum-rules-8-3-16.45124/

Funny. I just came back on here to say that I think the problem is actually being caused by a problem in the boot drive. I have been working from the console here, and after another reboot, I was working through the web interface, and I tried to access the replication tasks, and the drive threw an error on the console next to me and the web interface died.

So I have machine A, here, which seems to be having a problem with the boot disk (an SSD that's a couple of years old). It has 4 disks. One is the boot disk, and then it has 3 disks in raidz which are the data disks. I also have machine B, which is running the same version of FreeNAS on a system that is mostly identical except that the boot disk is different, not an SSD, and then minor config differences like the IP address and the replication tasks.

I don't want to just fail over to machine B because it looks like the replication tasks have been uncertain since this problem. They are supposed to happen every hour, but it looks like the last one was at 11:53 pm last night. So the data is not up to date.

Before this happened however, I was able to download and save the config from both machines A and B. My question now is, since I have this backup server that is running the same version of freenas, , can I take the boot drive out of the backup machine, put it in the primary, and copy the config from the backup on the zpool, or by uploading it from my backup on the system I'm working from?

If I do that, should I pull the raid disks until I have restored the config? The one thing that would be an unmitigated disaster would be damage the raid array.