Server CPU / Memory / Disk Read Spikes

[mag]

Running latest linux 64-bit server, 3.13.7. Not sure when this started but there are recurring spikes in CPU, memory, and disk read several times a day. These occur only when the server is running. I have tried running the server both natively and in the Docker container and see the same behavior on both. Nothing odd or usual about traffic to the server. Absolutely nothing shows in the server logs. Using the sqlite database.

Any ideas?

(Sorry some of the disk data got cut off in the screenshot but those are read spikes, no writes.)

[andr3pt]

Same occur to my server (3.13.7), as a new ts3 server or not, its the same…
Looks like I’m not the only one

[mag]

Who’s the right person to ping about this? @TS.ChrisR maybe?

[TS.ChrisR]

With given details we are not sure what this could be.

• What are your system specs?
• Are you sure there is nothing in the server log?
• Is this a server with a non default license (a licensekey.dat file exists in your server folder)?
• Can you reach our accounting service accounting2.teamspeak.com ?
• Is there any file transfer activity around these times?
• Have your turned off the weblist reports?
  Can be done In the server edit dialog in your TS Client or with the ServerQuery or WebQuery.
  https://support.teamspeak.com/hc/en-us/articles/360002711677-How-do-I-add-or-remove-my-server-from-the-Server-List

[mag]

• a single core alma 8 vps, 1gb ram
• not a word to be found either in ts logs or in system logs
• yes we have a license key (ancient npl)
• accounting2 connection stable with sub-5ms ping
• file xfer disabled - more properly, nobody has upload perms
• weblist disabled

[TS.ChrisR]

Are you sure this is caused by the TS server at all (even when this is the only bigger software running)?

• Can you see what software does cause this?
  • Have you checked this in real time when it happens?
• Maybe a backup or other service is running on that machine?
  • Or any other Scripts or ServerQuery doing something on the server.

So far we have no clue what else this could be. In most of all cases there is something spamming the logs when our server causes such load.

[mag]

On a hunch I decided to increase the server’s swap file, and the problem has gone away. If I had to guess based on what I’ve read, I would blame dnf. No hard data on this but it checks for data on a regular basis like the times when the spikes happen, and it has become more memory-hungry over time as the repos it reads data from have gotten bigger.

@andr3pt can you try this and see if this helps you?

[EGOiST1991]

Same for me here.

3.13.7 Linux 64bit spikes like crazy every 10 minutes. I noticed it on my consumption reading in home assistant.

Had to shutdown 11 lxc containers one by one double check its really teamspeak3 server.

Can confirm there is something wrong. Increasing swap storage did not help me.