Welcome to MilkyWay@home

Server Problems? - currently the U/L one

Message boards : Number crunching : Server Problems? - currently the U/L one
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 5750 - Posted: 1 Nov 2008, 14:30:15 UTC

Are there server problems as they are difficult to access or very slow to load?

I don't know if Milksop's fast client is helping towards the issues as crunchers run out of work and hammering the servers?

All my rigs are crunching MW fine and D/Ling work. But, the U/Ling of results and "waiting to report" seems to be particularly slow.

DI cannot see it in the Server Status, but do any of the servers need kicking?
ID: 5750 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
BarryAZ

Send message
Joined: 1 Sep 08
Posts: 520
Credit: 302,524,931
RAC: 15
Message 5751 - Posted: 1 Nov 2008, 14:35:56 UTC - in response to Message 5750.  

I think it is simply the effect of the fast client -- it pushes results and requests new work at something like 50 times the previous rate. Sort of a jailbreak effect on the servers. I think the problems will persist until folks at the project can release new larger (and slower) workunits to reduce the load on the server. It is one of the unintended but predictable effects of having the optimixed client available to all who wish to use it.



Are there server problems as they are difficult to access or very slow to load?

I don't know if Milksop's fast client is helping towards the issues as crunchers run out of work and hammering the servers?

All my rigs are crunching MW fine and D/Ling work. But, the U/Ling of results and "waiting to report" seems to be particularly slow.

DI cannot see it in the Server Status, but do any of the servers need kicking?


ID: 5751 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Dave Przybylo
Avatar

Send message
Joined: 5 Feb 08
Posts: 236
Credit: 49,648
RAC: 0
Message 5753 - Posted: 1 Nov 2008, 16:50:31 UTC

There were some quite serious corruptions of the "profile" table in the database. I have fixed them and now everything should be back to normal (including editing of profiles and user of the day).
Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute
ID: 5753 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Benjamin Rietveld

Send message
Joined: 26 Sep 08
Posts: 11
Credit: 17,597
RAC: 0
Message 5754 - Posted: 1 Nov 2008, 16:55:27 UTC
Last modified: 1 Nov 2008, 16:56:06 UTC

I've gotten all kinds of errors today, but at the moment I'm getting

01/11/2008 17:54:40|Milkyway@home|Sending scheduler request: Requested by user. Requesting 876 seconds of work, reporting 0 completed tasks
01/11/2008 17:54:45|Milkyway@home|Scheduler request completed: got 0 new tasks
01/11/2008 17:54:45|Milkyway@home|Message from server: Server error: can't attach shared memory

Also, the feeder and like half of the other stuff is offline?
ID: 5754 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
EigenState

Send message
Joined: 19 Apr 08
Posts: 13
Credit: 151,588
RAC: 0
Message 5755 - Posted: 1 Nov 2008, 16:58:13 UTC - in response to Message 5754.  

I've gotten all kinds of errors today, but at the moment I'm getting

01/11/2008 17:54:40|Milkyway@home|Sending scheduler request: Requested by user. Requesting 876 seconds of work, reporting 0 completed tasks
01/11/2008 17:54:45|Milkyway@home|Scheduler request completed: got 0 new tasks
01/11/2008 17:54:45|Milkyway@home|Message from server: Server error: can't attach shared memory

Also, the feeder and like half of the other stuff is offline?



Same problem here.
ID: 5755 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Dave Przybylo
Avatar

Send message
Joined: 5 Feb 08
Posts: 236
Credit: 49,648
RAC: 0
Message 5756 - Posted: 1 Nov 2008, 16:59:11 UTC - in response to Message 5755.  

Yup. They are back on as of 5 seconds ago.
Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute
ID: 5756 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile DoctorNow
Avatar

Send message
Joined: 28 Aug 07
Posts: 146
Credit: 10,545,421
RAC: 9,640
Message 5757 - Posted: 1 Nov 2008, 17:00:53 UTC - in response to Message 5753.  

(including editing of profiles ...).

Yeah, it works now! Thanx Dave! :-)
Member of BOINC@Heidelberg and ATA!

My BOINCstats
ID: 5757 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 5758 - Posted: 1 Nov 2008, 17:02:39 UTC
Last modified: 1 Nov 2008, 17:09:46 UTC

Dave

Thanks for coming in on a Saturday and give the servers a tickle and kick where appropriate.

I assume the Server status notice of zero WU available to send (when these servers/assimilators/etc are back online) is due to those of us using Milksop's fast client. This will suck down WUs at a tremendous rate.

Can the creation of new WUs be made faster?

As you say, the servers are working again and I have sent the WU results "ready to report". Now just waiting for some new WUs to crunch. Just need to streamline this with the other project I am now crunching as well.
ID: 5758 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Dave Przybylo
Avatar

Send message
Joined: 5 Feb 08
Posts: 236
Credit: 49,648
RAC: 0
Message 5759 - Posted: 1 Nov 2008, 17:40:18 UTC - in response to Message 5758.  

Yes, the optimized client is crunching units faster than we can churn them out. This might be a bottleneck and a downside to the optimized client. You folks have to battle for new workunits. However, I'll see what we can do, if anything, about this problem.
Dave Przybylo
MilkyWay@home Developer
Department of Computer Science
Rensselaer Polytechnic Institute
ID: 5759 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 5762 - Posted: 1 Nov 2008, 17:59:18 UTC

I wouldn't be too concerned at this stage. If the work is getting completed in 2% of the time it use to take, then I think we can wait for some work if necessary!

Hopefully more science is being done for the benefit of the project!
ID: 5762 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 5763 - Posted: 1 Nov 2008, 18:01:54 UTC - in response to Message 5762.  

I wouldn't be too concerned at this stage. If the work is getting completed in 2% of the time it use to take, then I think we can wait for some work if necessary!

Hopefully more science is being done for the benefit of the project!



Agreed

The now spare cycles are available to other distributed projects
ID: 5763 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Benjamin Rietveld

Send message
Joined: 26 Sep 08
Posts: 11
Credit: 17,597
RAC: 0
Message 5766 - Posted: 1 Nov 2008, 18:40:33 UTC

indeed, and even though the word generator can't keep up we're still doing way way way more crunching than we used to :)
ID: 5766 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Benjamin Rietveld

Send message
Joined: 26 Sep 08
Posts: 11
Credit: 17,597
RAC: 0
Message 5773 - Posted: 1 Nov 2008, 20:44:46 UTC - in response to Message 5750.  

I haven't got a workunit or seen work available for hours now. Is that just me or has the workgenerator crashed?
ID: 5773 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Cori
Avatar

Send message
Joined: 27 Aug 07
Posts: 647
Credit: 27,592,547
RAC: 0
Message 5774 - Posted: 1 Nov 2008, 20:45:46 UTC - in response to Message 5773.  

I haven't got a workunit or seen work available for hours now. Is that just me or has the workgenerator crashed?

I think with the fast apps we've crunched the current WU queue empty. ;-)
Lovely greetings, Cori
ID: 5774 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 5776 - Posted: 1 Nov 2008, 20:49:02 UTC - in response to Message 5774.  

I haven't got a workunit or seen work available for hours now. Is that just me or has the workgenerator crashed?

I think with the fast apps we've crunched the current WU queue empty. ;-)



That is my impression as well.

The fast client just sucks down the queue available and the Feeder just cannot keep up.

I think those with the old client, and a cache, are still crunching. But this will not last as there is no WUs to send even to them.
ID: 5776 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Cori
Avatar

Send message
Joined: 27 Aug 07
Posts: 647
Credit: 27,592,547
RAC: 0
Message 5777 - Posted: 1 Nov 2008, 21:02:26 UTC - in response to Message 5776.  

But since new WUs are created from results crunched before we may expect many WUs in the near future again (I hope). ;-)
Lovely greetings, Cori
ID: 5777 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Gavin Shaw
Avatar

Send message
Joined: 16 Jan 08
Posts: 98
Credit: 1,371,299
RAC: 0
Message 5778 - Posted: 1 Nov 2008, 21:08:07 UTC

As fast as they are created they are taken up by someone. I think it is now a matter of requesting more work at the right time to get more units.

Never surrender and never give up. In the darkest hour there is always hope.

ID: 5778 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
John Clark

Send message
Joined: 4 Oct 08
Posts: 1734
Credit: 64,228,409
RAC: 0
Message 5779 - Posted: 1 Nov 2008, 21:09:32 UTC - in response to Message 5777.  

But since new WUs are created from results crunched before we may expect many WUs in the near future again (I hope). ;-)


That is certainly a point worth considering.

The more we crunch now, with long gaps, the more we may have to crunch soon. But, it still comes down to the capacity of the feeder, which is set up for the outputs before the fast client came out yesterday.

It should be remembered that the Admins are tweaking a new client, which is based on a similar code to the Milksop client release. So, in the light of this per-new MW fast client release, they may need to consider what they want to study and how quickly.

There is discussion around that in other threads.
ID: 5779 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 5782 - Posted: 1 Nov 2008, 21:17:07 UTC - in response to Message 5779.  

It should be remembered that the Admins are tweaking a new client, which is based on a similar code to the Milksop client release. So, in the light of this per-new MW fast client release, they may need to consider what they want to study and how quickly.


I do remember a mention about doing a larger portion than is being done currently, so that should make the wu's longer again. (not sure if it will happen or not)
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 5782 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 5793 - Posted: 2 Nov 2008, 1:13:19 UTC

I don't think many will appear untill the next group is added. The total has gone from 36000 to 17700 wu's.
ID: 5793 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
1 · 2 · 3 · Next

Message boards : Number crunching : Server Problems? - currently the U/L one

©2024 Astroinformatics Group