Last week I got setup with a new dedicated server that is running 4x SCSI drives in RAID 10 on CentOS. The machine is pretty lightly loaded at the moment and has been running for 8 days.

The process 'scsi_eh_3' has a CPU time of 13:09.67 in top. This is 3x the next closest CPU time of 4:03.93 for mysqld. Is it normal for the scsi error handling process to use this much CPU time? (Seems like it goes to 0.3% usage ever 5 seconds.) Could this be an indicator that one or more of the drives is having issues?


Not a 100% answer but my scsi_eh_x processes on a couple of servers that have been up for a long long time show 0 cpu time, so I would say yes, get the diags for you scsi card and run them looks like you have something going wrong.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.