stress2 and results
Peter Holm
peter at holm.cc
Fri Jan 2 21:01:01 UTC 2015
On Fri, Jan 02, 2015 at 10:08:34AM -0500, Susan Stanziano wrote:
> Happy New Year, everyone:
>
> On Dec. 22, Andy Zhang asked some questions about stress2 which I also had but no one has responded yet so I'll ask mine:
>
> I've read the available doc (two .pdf files) but don't see how to understand results or where to find results. It looks like there are no
> created result files. Is this correct? So, is the goal to see a panic and then de-bug the panic?
>
Yes, that's right.
> I've run the suite on a VM here and am at Loop #1118 now (about 48 hours). The system gets low on resources at various times but hasn't panic'd yet.
>
> What is the accepted way to interpret this situation? I would like to make this suite a useful part of our testing effort here, but I need
> to undestand the intended goals.
>
The primary goal is to crash the kernel. If running the tests for a
long time does not cause a panic, that is often interpreted as
"success". A different view is that the test suite is broken :)
Look at the (regression) tests in stress2/misc, and run them by
./all.sh -o
This should take some 36 hours on modern hardware. Remember to add
a swap disk, as many of the tests rely on VM pressure.
> Thanks,
> Susan Stanziano
> _______________________________________________
> freebsd-testing at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-testing
> To unsubscribe, send any mail to "freebsd-testing-unsubscribe at freebsd.org"
--
Peter
More information about the freebsd-testing
mailing list