PERC5/E and PERC6/i oddities with OpenSuSE 11.0
jansen at astro.wisc.edu
Sat Feb 21 06:44:51 CST 2009
We have a PowerEdge 2900 as a head node to a 72 node cluster we
just built. The 2900 has an internal 5TB RAID array controlled
by the integrated PERC6/i and two 10TB MD1000 RAID arrays controlled
by two separate PERC5/E cards. We are monitoring the arrays and
temperature with OMSA 5.3. We originally installed OMSA 5.5 but
the arrays always showed up as "degraded". This all seems to work
well with the follow two exceptions: 1) we frequently see the message
"kernel: megasas: Failed to copy out to user sense data" in the
system logs. This does not *appear* to be causing problems
but it doesn't give me a warm and fuzzy feeling, 2) the process
"dsm_sa_datamgr3" is using around 100% of one of the CPUs.
The operating system we are using is OpenSUSE 11.0 x86_64, kernel
version is 22.214.171.124-0.1-default. The version of the megaraid_sas
module appears to be 00.00.03.16-rc1.
Does anyone have any clues about the cause of these two problems?
More information about the Linux-PowerEdge