VPS Randomly Hanging

[ljj3]

One of our VPS's - perfectly behaved for 5 years has started totally hanging every few hours. Loading is minimal, there is no DOS attack, absolutely no errors. Just freezing which can only be restarted through vendor control panel (SSH frozen)

Sister VPS on same hypervisor running perfectly.

Only difference is hanging VPS has KernelCare.

Vendor suggests restoring from backup prior to start of issue.

Any ideas???

Thanks!

 

[ffeingol]

Anything in /var/log/messages?

 

[ljj3]

Nothing! Just the usual stream of firewall messages.

 

[ljj3]

Also - no resource issue, nada, happens in. middle of day or middle of night.


very lightly loaded system...

 

[ljj3]

UPDATE: Server restored with snapshot prior to problem and....

It hung again - no errors just hung.

 

[cPRex]

If the system is hanging even after a restore from before the problem happened, it seems like an issue with the parent server. You said other VPS machines on the same hypervisor are running normally - if that from data you have, or is that what the host told you? Ultimately, you'll have to work with the host, as there isn't anything in cPanel that will make a server freeze up like that.

 

[ljj3]

From what I can tell, the other VPS which actually has a much higher load is doing fine. They "claim" its not an infrstructure issue but really, what else can it be - if the vps logs no errors it must be a problem above it, right???

 

[cPRex]

Correct

 

[sparek-3]

What Hypervisor is being used for this VPS?

What OS is the VPS is using?

What OS is the parent node using?


I've had an issue with KVM AlmaLinux 8 VPSs running on an AlmaLinux 8 parent node, freezing up for no apparent reason. Issues been going on for 6 months or better. I've found absolutely nothing to diagnose the issue. The VPSs just freeze up every 2 or 3 weeks or so. The VPSs essentially have to have the power unplugged in order to restart. I'm getting at wits end trying to diagnose the issue. Doesn't help that there's no patterned timetable of these incidents.

If you've been running this VPS for 5 years, I kind of doubt you're using AlmaLinux 8 - but your symptoms sound similar to mine. I don't have any resolution.

 

[ljj3]

Centos 7 on the VPS. - this started 9/3 which makes you wonder if they changed something over the weekend.

 

[ljj3]

Hangs have gone down to 2 a day - at no specific time. Vendor believes it is DDOS activity - the console is a steady stream of firewall log messages, but so are all my other servers. They sent out a notice that the hypervisor we're on is going down for maintenance next week.

Thoughts?

 

[cPRex]

If it's DDoS activity, it doesn't seem like it would be related to your particular system with the details you've posted. Watching "apachectl status" would show the scoreboard and you can see if you're getting hit, but it still doesn't sound like an issue with your machine to me.

 

[ljj3]

Current Time: Thursday, 07-Sep-2023 11:32:06 EDT





Restart Time: Thursday, 07-Sep-2023 00:51:44 EDT





Parent Server Config. Generation: 1





Parent Server MPM Generation: 0





Server uptime: 10 hours 40 minutes 22 seconds





Server load: 0.21 0.27 0.22





Total accesses: 11295 - Total Traffic: 583.0 MB - Total Duration: 9401975





CPU Usage: u43.15 s12.78 cu0 cs0 - .146% CPU load





.294 requests/sec - 15.5 kB/second - 52.8 kB/request - 832.401 ms/request





2 requests currently being processed, 123 idle workers

 

[cPRex]

An there you go - definitely not your system being DoS'd, or at least not right now.

 

[ljj3]

Thanks for the sanity check!