Server Problem
 

After some kind of crash in the night there are some table files brocken on the server. I have stopped all of the stats and replenishement jobs.
The queues might run empty before i can repair it.

On Sierpinski the job allocation should continue to run and the transfer to the main table would also work but to make sure not to damage even more data i have stopped that part also.

On riesel the problem is within the job allocation queue. So i fear the will be no work allocated correctly.

I have to go to work now so till tonight there is nothing i can do to solve the problems.

Sorry,

Lars

For sierpinski the llrnet server runs without problems and we should have enough work for at least 2-3 days before running empty if we continue to work at the current rate.

For riesel side i am very confident to have at least the llrnet server up again this evening as it is very easy to create new tables if there is no chance to repair the original ones. The only problem would be that there is a chance that about 100 WU would have to be done again with a slide chance of not beeing credited.

Lars

Status update:

good news so far.

It looks like the Table repair worked without any lost data.(which i have to crosscheck later today)

Riesel side started to hand out new WU at 15:50 CET.
There is enough work in all queues that i can do all needed crosschecks before i activate the sheduler and the stats.

Most important:

[B][SIZE="7"]NO WORK LOST :grin: :grin: [/SIZE][/B]

Lars

Now SR5 is also back on track.
For both parts the sheduler and the stats engine is up again.

Lars

thanks lars for your fast answer to this crash !

No work lost :smile:

BTW
I was having a look through one of the client logs whilst the server was misbehaving and found this
33448*5^228454+1 is a probable prime.

sierpinski server is down since 21 mai @ 01:30

[QUOTE=tnerual;106583]sierpinski server is down since 21 mai @ 01:30[/QUOTE]

seems to be solved since 8:30 ...