WU completed 100% but still running for several hours

[candido]

I have this WU that has reached 100% completion after running for 292 hours.
Since then it has continued to run and has been doing so for 12 hours.
Is this normal?
Should I abort? (I wouldn't like to do that after having been running this WU for 15 days)

WU is cms_GA-p[b-Lin64f-2]_1_Vibrio-vulnificus-YJ016_BA000038.cir.EMBL_RF00028_Intron_gpI_1358679723_2095

thanks

[ClaudiusD]

Yes its normal and please dont abort the WU. Some WUs takes more time. In the baddest case, you can double the time, but often they are "real" finished after 20-30% more time - perhaps in 5days for you.

[candido]

I see.
I'll continue to run the WU.
Many thanks.

[ClaudiusD]

No problem and welcome to the forum!

If Boinc means your WU has not enough time to complete, you have to post it into the forum (here). At the moment some WUs have to be prolonged. If you have any question, feel free to ask

[ChristianB]

I extended the deadline on the server for this WU.

[candido]

I have another 4 tasks (Virtual machine tasks) in another computer (a i7-2630M processor).
do these virtual machines task have the same behaviour or do they end at 100%.
some of these tasks have an estimate of 2000 hours, which is a tremendous amount of processing time on a i7...
thanks

Edit: they seem to jump to 98.765% complete after the first checkpointing.
By the way, checkppoint is a huge improvement!!

[ClaudiusD]

This project is still in beta and has some little problems.

Normally the VM-WUs will extend themself. And normally they should display an estimated finish-time.

For the prolongation (is this the correct word?) of the WUs one of the admins will do it. I think ChristianB does this. So you have normally no problems with the length of the WU. But at the moment there could be some problems on some machines.

My own i7 runs with 8 VM-WUs and had no problems so far.


The total length of the WUs can be abnormous big. The problem was the checkpointing... A blackout after 1500hours is fatal!
Look at these WUs: Monster Hunter. The longest was about 2000 hours (without a single restart of the computer). This ultralong WUs are hard to "crunch", so these will be surely the first on a virtual machine!

The other problem - the 98.765%... seems to be normal (and they are looking for the problem)

My VM-WUs with 98,765% (i had some WUs before from the first "round" - and they showed no problems):

[yoyo]

The VM Workunit sends so called trickle messages to the server. With this the server know that the workunit is still active and if it gets close to it's deadline the server will extend the deadline for this task. If this is properly done you can see on the deadlines seen on the server for your running tasks.

yoyo