This is Lite Plone Theme
You are here: Home System Status COSMOS at-risk unexpectedly [UPDATED 20:43 UTC]

COSMOS at-risk unexpectedly [UPDATED 20:43 UTC]

Dear users,

All systems are now back online and ready for jobs.

Apologies for the inconvenience caused today.

Kind regards, COSMOS Management

Dear users,

We had to cancel two jobs, Moab.232258 and Moab.232206. We are very sorry for this. The other jobs finished in time before the shutting down of CXFS but these had so many wall-clock hours still left that it was not practical to have everyone else wait for just two jobs to finish.

We are sorry for the inconvenience caused.

Best regards, COSMOS Management

old content

Dear users,

Universe decided to start misbehaving when the CXFS was turned off on it, so we now think it is safest that we reboot every system; we will also take this opportunity to install a recently released update from SGI which fixes one of the problems cosmos2 has been experiencing of late (the problem exists on all systems, but the others have just been lucky not to suffer from it).

We are sorry for the long and inconvenient and more inconvenient than we led you to believe interruption. We will update you when things are back to normal.

Best regards, COSMOS Management

Older content:

Dear users,

It turns out the problem with /fast/space is more serious than we thought and will require every since CXFS filesystem to be taken off-line to be fixed.

We have turned all the queues off and we will have to kill any process still accessing CXFS when we eventually want to take them off-line but otherwise we should be able to keep the systems up. We will update you if we need to reboot them.

Best regards,
COSMOS Management

Older content:

Dear users,

As some of you will have noticed, /fast/space has become inaccessible. We do not yet know what caused this nor how big a disruption we will need to impose on the machines, but please consider the systems at-risk until further notice.

The systems are available (but /fast/space is not).

The systems may be rebooted without notice until this has been resolved.

We are sorry for the inconvenience caused.

Best regards,
COSMOS Management