1

I have an HP ProLiant DL380 G7 server, with Intel Xeon. I also have it with RAID 1 level on its HDD's. It has Ubuntu 16.04.3 LTS running on its SSD's.

Now, my problem is that sometimes power surges cause my server to crash, and when it comes back up some processes are not initialized. This I can tell by checking my processes with ps and similar tools and see they aren't running after the crash, for which I then have to start manually.

I have these processes scheduled on my crontab to start @reboot, and they do come back up when doing a "proper" reboot.

What could be the cause for this processes not starting? Is there something I can do to overcome this and be able to start them automatically?

Some other important information I can provide is that I suspect that it may have something to do with its power sources, as I have all my rack connected to a UPS, but only this server gives me that problem.

Could this contribute to the problem I am facing? I have seen that when the surges happen, some Amber LEDs start blinking for a moment (the power source ones) before returning to normal conditions.

First time asking here, so if you need any other information do tell me so I can provide it :)

DarkCygnus
  • 131
  • 5
  • What do the logs say? Cron should log its attempts to start the services. – Jenny D Dec 02 '17 at 10:15
  • Also, I can't help wondering why you have them set to be started by cron @reboot instead of having init start them... – Jenny D Dec 02 '17 at 10:15
  • Also, after a "crash" (as opposed to a scheduled reboot), does the server do a complete reboot or is there something else that happens? Again, check logs! – Jenny D Dec 02 '17 at 10:16
  • Hey @JennyD I'll check the cron logs, thanks for the suggestion :) about cron I found it more... convenient? Than init, would init not show this problems? And yes, the server then reboots on its own, just like that, but I see the LEDs as mentioned on the post. I'll check the logs first thing in the morning Monday thanks again – DarkCygnus Dec 02 '17 at 16:35
  • 1
    Generally speaking, you'd use init/systemd to start daemons that should keep running, while cron is for things that should run once at reboot and then not run anymore. Init has better logging when stuff breaks and better ways of handling dependencies (which may be the issue here). In this case, I'd also suggest improving the logging from the scripts starting the processes and/or from the processes themselves. – Jenny D Dec 02 '17 at 16:49

0 Answers0