For no obvious reason mpt2sas started to spam my system logs on a Wheezy box with an LSI SAS 9200 controller, no raid function, many disks attached) and I worried about something being broken or breaking right away. Just for the record: I would recommend this particular controller as being rock-solid and fast under heavy load.
mpt2sas0: log_info(0x30030101): originator(IOP), code(0x03), sub_code(0x0101)
Sadly, there was no useful documentation from LSI and no good hints anywhere else. Also the lsi_log_decode i stumbled upon had no clue at all. What led me to the cause of these messages was an old bug report from CentOS which finally mentioned the exact log message i got and pointed me to an issue with calls to non-existent pages from a controller related management software.
I remembered testing different cli tools. One of these was sas2ircu respectively sas2ircu-status which didn’t meet my requirements while testing, but still was installed on the system. Apparently sas2ircu-status was trying to request some information from the controller to which it responded by putting broccoli in his ears. As sas2ircu and sas2ircu-status anyway are of no use to me, the solution was to simply purge the sas2ircu* packages and hey presto, log messages stopped immediately.
Thanks very much for posting this, I recently ran into this message myself and removing sas2ircu (which check_raid was invoking automatically) fixes the issue. In my case this was a cross-flashed adapter in IT mode (originally IR) which has no RAID.
Danke Alex, Guide ist immernoch aktuell.
Top
For the record, I’ve stumbled across this in my kern.log:
mpt3sas_cm0: log_info(0x30030101): originator(IOP), code(0x03), sub_code(0x0101)
As with Chris, it’s an adapter with IT instead of IR firmware (so, acting as HBA), namely the LSI SAS3008 onboard a Supermicro server mainboard.
Also as with Chris, this was caused by a monitoring check, which was calling
sas3ircu 0 DISPLAY
to get a list of all disks. Whenever this command is issued, the log line appears for me.