Does aborting a task make it immediately available for re-issue - or thereabouts?

Pretty much. As long as work unit hasn't reached its "Max # of errors" total, which is usually 3 for the work units being sent out nowadays.

Thanks, I thought that was probably the case.

Following the recent outage among the work units I have got are some from earlier batches - 660, 709 and 719. All are on their third go. Should I allow these to run or should I abort?

Nothing in the 700 range has been closed yet, so those 2 are OK.
I don't see 660 in the closed list either, but it is getting a bit old. Best to get it run soonish.

Yesterday I completed a 647 that one of my computers picked up 12 hours before the nam50's came out, so good to get that out of the way.

Which leaves a 719 still running with 2 days to go. Just under 8 days run time so far to reach 85%, if that helps.

Thanks Les. The 660 is about 25% complete after just under 2 days so shouldn't take much longer. The 709 and 719 haven't started yet. The 735s that I have had have been taking between 3 nad 4 days to complete. I'll keep plodding.

Hi folks,
I got a hadcm3s_a18n_203412_120_599 on my Linux box. It is in its 3rd attempt and I wonder should I let it finished or this batch is of no scientific interest any more?

I got a hadcm3s_a18n_203412_120_599 on my Linux box. It is in its 3rd attempt and I wonder should I let it finished or this batch is of no scientific interest any more?

It is still in the current list, not in batches to be closed or closed lists. But as you note it is quite old.

But it's a "short", which are usually a challenge to complete, so you could have a go at it just to see if your computer can do it. :)

I ruined it. I did system update before the WU finished and after restart it ended with computation error. Here is the error I updated from 14.04 to 16.04 LTS and after the crash I checked missing libraries and I had to install gcc.4.7-multilib. Not sure if this was the reason, but I really had to let if finish first - it was its 3rd attempt