Hi.

We're running a heavily loaded SunRay system on RHEL4 (yes, unsupported
but still).
Once every week or so, a server crashes due to a null pointer within
utadem. More specifically this happens in the oss_ioctl routine while
executing a SNDCTL_DSP_GETODELAY ioctl.

We're suspecting this is caused by some kind of race condition, the
servers are SUN V40z machines having a total of 8 execution cores, so
there are lots of possibilities for real racing problems if the code
isn't right.
I'm not sure this may be due to the fact that we're running this on
RHEL4 instead of the supported RHEL3. Could there be significant changes
between these systems that make the utadem code unfit here.

Has anyone seen this at all, on any systems? Any clues would be helpful.
I really have to pick this bug out, as there are lots of users who are
affected.

And yes, we know we shouldn't do this, but there are lots of surrounding
requirements that led to the decision to launch this into real
production anyway. Hoping to see SRSS3.2 soon though.

Sincerely

Christer Bernérus
IT services
Chalmes University of Technogy
Gothenburg, Sweden

_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users

Reply via email to