• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Issue Server hangs up every day

LuaNomed

New Pleskian
Hello,

I have a very strange issue.
About three days ago I updated my Plesk to Obsidian from Onyx and my Debian 8 to 9. Everything ran fine.
The server was up and running, the domains were working. I also had to get a new key from my reseller as my old one still was one from the old licensing system. But installing the new Web Pro Edition Key also worked fine.

Ab bit later, I am not sure how long exactly, but it was always in the evening for the next 3 days in a row, the server acted strangely. I couldn't reach any of the domains and after I logged into Plesk I got the message "The license key is invalid" This alone would let me turn to my provider how gave me the key, but the page also reloaded the whole time, like every second. So I was unable to put in the key again or any new one.

I then decided to go the SSH route an did a plesk repair. This solved the issue. Domains were back online, key was valid again, I could log in to Plesk. About 24 hours later the same game again. And now for the third time. It worked an hour ago or maybe two, so it definitely worked for 19-20 hours this time, not sure for the other two days.

Anybody an idea, how to fix that? I am so glad it is only a backup server with some private domains on it, but I need to update my client-server, too, and I am now a bit freaked out. The client-server already has a new key from the new licensing model, it is working fine there. But this server is still Onyx with Debian 8.

P.S: I forgot, before I could login via SSH I had to reboot the server ones over my provider's rescue system as even SSH was failing "The SSH session has terminated with error. Reason: FlowSocketReader: Error receiving bytes. Windows error 10054: Eine vorhandene Verbindung wurde vom Remotehost geschlossen." After the reboot the plesk login still had the invalid key and constant reload error, but SSH was working again. The domains also worked again after the server reboot.
 
Last edited:
So I investigated a bit further and rewrite the hole post, as got a bit messy.

-----------------

First thanks IgorG for pointing out, that the problems with the thread was only on my side. Still took me multiple tries to get it back working, never had such issues with a forum software before. Still not able to access the thread on my main firefox profile, but I created a new one that worked.

Back to the actual issue. As I wrote the key is fine again after the plesk repair, there were no issues when I installed it first and I reinstalled it, too. Netherless I supported Licensing and Purchase support but I am very sure there is something else going on and the key thing is just a symptom.

When the server had the problems today I was greetd by a mysql error when I tried to log into plesk. It was gone after a reload and I don't remember what it said. But new issues also showed, when I tried to login via SSH :"/root/.rbenv/bin/rbenv: fork: Nicht genügend Hauptspeicher verfügbar" (engl: Not enough main memory available.). Tried a plesk repair over SSH, but plesk is not found. Only the restart of the server through my providers login helped.

After the restart I was able to get back into plesk and checked the monitoring, everything there had free resources over the last days, so don't know what the message memory means. But I noticed something else, i did a ps axv a few minutes ago while the server is running okay, there are 5 instances of apache2 (/usr/sbin/apache2 -k start), when I did the command before I did the server restart, there were arounf 30-50 apache2 instances, which didn't seem normal, or? It also shows under monitorin, that the count of sp_state_sleeping goes from around 130 to 260, it stays at that high value all the time the server is struggling and after the restart back to 130. It takes around 6 hours to go from 130 to 260. Is that normal? Could that cause the strange behavior of my server?
 
Last edited:
I found some new symptons and rewrote the post above. Maybe somebody has a clue, this is hopefully a bit more helpful then the invalid key issue.

And another update.
My fresh installed test server with Debian 9, Plesk 18.0.20 via Autoinstall just got the same invalid key error.
It did run fine till 5 minutes ago.
I only have one domain on this server, there was nothing on that domain till today. I had activated ruby for the domain, but I only installed the app 30 minutes ago. I did a pull from git to get the app on the domain, then I did run the bundle installer but got errors, as I did forget to install some packages needed for compile. I decided to install the ruby developer stuff from plesk to see if they have everything included I need. They installed without problems, after closing the updater to get back to plesk, the invalid key messaged poped up.

After reinstalling my key, the message is gone for now. I did send my provider a mail, if they have an idea what the hell is going on.
I have 4-5 servers, all with their keys, never one single problem and now 2 servers right after update to Obsidian. This release seems really buggy.
 
Last edited:
Back
Top