0

I have a problem that I cannot wrap my head around. One of our machines (physical dedicated server, Debian 10, nginx, mysql, PHP 7.3) is running a php application that also uses the php-cli regularly through some cronjobs.

The cronjobs do nothing special, just some file system work and a few mySQL requests.

Yet, the machine reboots every other hour without any log file entry (nothing in syslog, kern.txt, dmesg). The SSDs are fine, RAM was tested successfully without faults.

We disabled the cronjobs now, and the machine has been up for more that 12 hours now.

Originally, I was suspecting hardware issues, because I never heard that a php instance could completely kill off Debian. That would mean that a php script could cause a kernel panic. Is this at all possible?

And if so, is there any way to debug this?

2
  • 1
    check in cron logs, mysql logs, enable crash dump and journeld.
    – asktyagi
    Commented Mar 28, 2020 at 9:44
  • Nothing in the logs, will install kdump-tools - thanks for the heads up!
    – Stimpy
    Commented Mar 28, 2020 at 10:10

1 Answer 1

1

This is conceivable, but highly unlikely. You mention "file system work" - if this is being dine through system commands it is technicallt possible, but not something that should be happening without your knowledge.

I recently got a vm server running multiple web servers which was crashing every day or so - not as frequently as hours, but similarly - no logs or anything. I never got to the bottom of it, but the identical VMs all worked fine on another server. I suspect the cause was a bad BIOS update/fix/workaround for spectre/meltdown, bad memory or a bad disk controller - unfortunately I will never know which as I returned the server.

2
  • Thanks for the input on the possibilites! The server was running fine for about a year now. Last resort for us now is to do a 24h stress test, which should rule out any bad memory / disk controller issues.
    – Stimpy
    Commented Mar 28, 2020 at 10:04
  • 1
    I doubt stress testing can rule out disk controller issues (which is not the same as saying there are controller issues). By the same token though, absent firmware updates these are not likely a cause if the system was running OK for a year.
    – davidgo
    Commented Mar 28, 2020 at 10:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .