I’ve recently encountered a situation with the CD-ROM drive on one of our Xen dom0s which led to significant spikes in the CPU load averages. /var/log/messages had several entries which stated “kernel: hda: drive not ready for command”. I did a search and found a thread at Error Message - hdd: drive not ready for command [Archive] - Ubuntu Forums with the same problem. I did what one user had suggested which was to attempt to mount the device. This seemed to break a loop which was running and caused the load averages to immediately drop.
Interestingly, and thankfully, Xen does an outstanding job of segregating problems like this. As a result, none of our domUs were affected by the problem. They all chugged along as if nothing were wrong.
Has anyone else encountered this? Does anyone know what causes it and if it is preventable?
Leading up to the error, only this is displayed:
Jun 20 12:00:01 xenhost4 /usr/sbin/cron[28550]: (root) CMD ( /usr/lib64/sa/sa2 -A)
Jun 20 12:29:29 xenhost4 syslog-ng[6947]: STATS: dropped 0
Jun 20 13:29:29 xenhost4 syslog-ng[6947]: STATS: dropped 0
Jun 20 14:29:30 xenhost4 syslog-ng[6947]: STATS: dropped 0
Jun 20 14:38:59 xenhost4 kernel: hda: irq timeout: status=0xd0 { Busy }
Jun 20 14:38:59 xenhost4 kernel: ide: failed opcode was: unknown
Jun 20 14:39:04 xenhost4 kernel: hda: status timeout: status=0xd0 { Busy }
Jun 20 14:39:04 xenhost4 kernel: ide: failed opcode was: unknown
Jun 20 14:39:04 xenhost4 kernel: hda: drive not ready for command
I’d be inclined to think that sa2 is attempting to access the CD-ROM drive which isn’t mounted but it ran cleanly for days prior to the problem.
Any help will be greatly appreciated.