My hard drives, connected to Ubuntu server are producing the following log every exactly 5 minutes.
Nov 1 14:10:50 localhost kernel: [ 1602.884936] ata2.00: hard resetting link Nov 1 14:10:51 localhost kernel: [ 1603.226804] ata2.01: hard resetting link Nov 1 14:10:52 localhost kernel: [ 1604.274533] ata2.00: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Nov 1 14:10:52 localhost kernel: [ 1604.274548] ata2.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Nov 1 14:10:52 localhost kernel: [ 1604.356669] ata2.00: configured for UDMA/133 Nov 1 14:10:52 localhost kernel: [ 1604.375247] ata2.01: configured for UDMA/133 Nov 1 14:10:52 localhost kernel: [ 1604.375265] ata2: EH complete
I don't think this is related to hard drive failure, because it happens for ALL hard drives connected and ONLY when I write
spindown_time = 12
/etc/hdparm.conf. The reason I add this value is to put disks into standby mode after 60 seconds, which is happening after that period (checked with hdparm -C).
The first clue I thought that smartd was running and spinning the drive. However, I couldn't find it in
ps -aux | grep smart.
Additionally, iostat does show that nobody accessed those drives, since Blk_read, Blk_wrtn remain unchanged. I also killed all processes that may be doing something with hdd(eg SAMBA). So I guess the problem is solely with hdparm...
I have no more clue where that 5 minute value hides.