Bulletin 192 - 2009 October 8

1. CSIRO cluster (burnet)

1. CSIRO cluster (burnet)

Users will have noticed the almost complete absence of service on burnet this week.

Both of the file servers continue to be unstable, so we have suspended all batch jobs for the time being.

We have moved away from the GPFS as a means to provide file services, since this had been previously unstable, and have gone back to the more widely used NFS services. However, we have been unable to establish a stable service under load, despite passing hardware diagnostic tests. The systems are just ceasing to function, and no error messages are being reported.

We are establishing another file server, using an alternative operating system environment and expect to have this operational today.

Please check the traffic lights at http://intra.hpsc.csiro.au/ or http://www.hpccc.gov.au/hpccc/system_status/ for updates.

We can get intermittent access to the files under the current circumstances: we can assist users to copy their files elsewhere. We will work towards providing access to other CSIRO ASC computing resource for users if the instability continues.

We apologise for the inconvenience caused.

[ page top ]