Bulletin 178 - 2008 June 26

1. System Outages

Note: "CSIRO" items can apply to BoM users of cherax and burnet

Due to maintenance on the chillers for the BoM computing facilities, there will be large scale outages on two occasions in the next fortnight. Cherax has additional scheduled outages for Wednesday evenings until the upgrade is complete.

All running jobs will be terminated at the necessary times. In order to avoid unnecessary inconvenience, please ensure that jobs which should not be restarted are marked as such. Torque and nqsII use the same directive for this:

#PBS -r n

The default is for jobs to be restarted when the system returns to service. You can change the restartable status of jobs that are already queued with 'qalter'.

On cherax and burnet, jobs that would run into the outage times may not be started. You may need to submit shorter jobs to keep using these systems in the times leading up to the outages.

Outage 1.

Eighteen SX-6 nodes (5 CSIRO, 13 Bureau) will be turned off leading up to 6.00am on Monday 30th of June. Non-operational workload will be reduced on the SX-6 nodes from 4.00am onwards. Non-operational jobs that cannot be held at 6.00am, will be deleted.

Cherax will be unavailable from 06:30 on Monday the 30th of June.

Salacia, Burnet and Nelson will be unavailable from 4:30pm on Friday the 27th of June.

We anticipate that availability will resume at around 2:00pm on Monday the 30th of June.

Outage 2.

Eighteen SX-6 nodes (5 CSIRO, 13 Bureau) will be turned off leading up to 6.00am on Wednesday 2nd of July. Non-operational workload will be reduced on the SX-6 nodes from 4.00am onwards. Non-operational jobs that cannot be held at 6.00am, will be deleted.

Cherax will be unavailable from 6:00am on Wednesday the 2nd of July.

Salacia, Burnet and Nelson will be unavailable from 4:30pm on Tuesday the 1st of July.

We anticipate that availability will resume at around 2:00pm on Wednesday the 2nd of July.

Resumption of service will be sooner if possible and later if necessary.

Apologies for the outage, and thank you for your patience.

[ page top ]