[Bi-users] Problems with the network on Bi
Kent Engström
kent at nsc.liu.se
Mon Jun 11 09:25:00 CEST 2018
Mats Kronberg <kronberg at nsc.liu.se> writes:
> There is an ongoing problem with the IB network on Bi (and Frost)
> which currently leaves the system unusable in practise. NSC is
> searching for the root cause of this and will update you when we have
> any new information.
Update: Bi is back since ~04:20. The problem seems to have started
before midnight (earliest alarm ~23:19).
Most running jobs have failed during the problem period.
The incident looks very similar to what happened February 21st, with a
hung Infiniband switch causing an outage of the Bi InfiniBand network
(not the same switch this time though).
We apologize for the inconvenience,
--
Kent Engström, National Supercomputer Centre
kent at nsc.liu.se, +46 13 28 4444
More information about the Bi-users
mailing list