Jess, Yes...it is a bug in the qus (Qlogic Ultra3 SCSI) HBA driver that wasn't draining all the IO from it's queue out to the disk. The work-around was to set the SCSI queue depth to 64. I have an IDR in place that has been working well. I am hopeful that a real patch will be released soon.
--Brett On Thu, May 28, 2009 at 8:41 AM, jess <jess at mlb.com> wrote: > Did Sun provide an explanation or solution for this issue? We have > experienced a similar process bug with nfsd on S10. > -- > This message posted from opensolaris.org > _______________________________________________ > mdb-discuss mailing list > mdb-discuss at opensolaris.org >
