1

I have noticed a sequence of this messages in my logs:

Oct 27 02:20:38 kernel: sd 0:0:0:0: [sda] Synchronizing SCSI cache
Oct 27 02:20:38 kernel: sd 0:0:0:0: [sda] Stopping disk
Oct 27 02:20:38 kernel: sd 0:0:0:0: [sda] Starting disk
Oct 27 02:21:13 kernel: sd 0:0:0:0: [sda] Synchronizing SCSI cache
Oct 27 02:21:13 kernel: sd 0:0:0:0: [sda] Stopping disk

From this, I understand that every minute my disks are waking up. Because this happens in a server that is very poorly used, I'd like to reduce the number of times my disks are waken up to the minimum. Does anybody know any way to do that? Regards.

chronos
  • 135
  • 5
  • If you are accessing it that often don't put it to sleep until a significant amount of time has passed – Ramhound Oct 27 '14 at 20:18

1 Answers1

0

you may want to install laptop mode tools, which would handle that (and other possible power savings).

Spinning down every minute is most definitely bad, you'll destroy your disk very soon

Matija Nalis
  • 2,616
  • 16
  • 24
  • "spinning down every minute is bad".. even with SSD, if there is (up to my knowledge) nothing spinning up? – chronos Oct 27 '14 at 21:29
  • @flix there is no "spinning down" on SSD (as nothing is "spinning"), so it can't be bad. Same ATA/SCSI commands might be used for SSD to speed up entering the lower-power sleep mode (instead of waiting for built-in timeout), though - and it should not negatively affect SSD life. But it won't bring you power savings like spinning down classical disks, either. – Matija Nalis Oct 27 '14 at 21:41
  • So, if there is nothing to be worried about with the messages, is there any way to stop flooding my logs? – chronos Nov 29 '14 at 09:01
  • @flix probably, depending on your logging daemon. For example, if using rsyslog, you'd probably want to [filter by text](http://www.rsyslog.com/discarding-unwanted-messages/); and for old syslogd, you'd need to ignore whole kernel.info facility/severity class. – Matija Nalis Dec 01 '14 at 20:10