Advanced search

Message boards : News : Server disk full being addressed

Author Message
Toni
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 9 Dec 08
Posts: 959
Credit: 4,353,973
RAC: 0
Level
Ala
Scientific publications
watwatwatwat
Message 53986 - Posted: 22 Mar 2020 | 20:48:44 UTC

I'm working to resolve. Thanks for the support :) -T

Toni
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 9 Dec 08
Posts: 959
Credit: 4,353,973
RAC: 0
Level
Ala
Scientific publications
watwatwatwat
Message 53998 - Posted: 23 Mar 2020 | 7:14:55 UTC - in response to Message 53986.

I'm suspending the servers to give time for the disks to empty on the main storage.

C270320
Send message
Joined: 27 Mar 20
Posts: 1
Credit: 2,223,415
RAC: 0
Level
Ala
Scientific publications
wat
Message 54123 - Posted: 27 Mar 2020 | 21:05:08 UTC - in response to Message 53998.
Last modified: 27 Mar 2020 | 22:04:00 UTC

Excuse me for asking - I am a complete, total, 'newbie' to shared computing.
Deleted original question was if "disk full" was causing my "request defferred".

All seems OK now: I am already 30% done with a new task (my 1st GPU task ever).

Thank you!

Toni
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 9 Dec 08
Posts: 959
Credit: 4,353,973
RAC: 0
Level
Ala
Scientific publications
watwatwatwat
Message 54210 - Posted: 3 Apr 2020 | 15:33:05 UTC - in response to Message 54123.
Last modified: 3 Apr 2020 | 15:34:52 UTC

Yes, usually server errors such as this are transient, and clients recover (sooner or later).

Also, your host is working just fine, not a single error :)

pball1224
Send message
Joined: 14 Jun 17
Posts: 1
Credit: 88,358,707
RAC: 96,348
Level
Thr
Scientific publications
watwatwat
Message 54214 - Posted: 3 Apr 2020 | 16:39:00 UTC

Task downloads appear to be bottlenecked. I'm seeing 0.13 and 0.08 KBps on two tasks my client has been trying to download for about 40 minutes now. Not sure if this is a problem, or just a massive amount of clients trying to refill their work queues after the outage.

djkra
Send message
Joined: 27 Feb 20
Posts: 2
Credit: 10,530,414
RAC: 0
Level
Pro
Scientific publications
wat
Message 54330 - Posted: 15 Apr 2020 | 2:35:59 UTC

wow , was gunna ask if gpugrid is working as i still wanna help. built a new super rig w this gpugrd in my target, so dissapointed to have to add other projects. dont wann join folding, stay loyal to something.. but had to add projects to use anything, reading the threads.. lots of negativaty. so much. when we all trying to help. so silly . i see a server status, but even after all this searching (45 min) i can only guess the server is getting hammered by all the new useres wanting to help. but idk. want to know if theres any update. is the server handing out units ?, or is it me., but i dont wanna get yelled at for asking.

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 209
Credit: 4,490,828,031
RAC: 1
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 54332 - Posted: 15 Apr 2020 | 5:08:41 UTC - in response to Message 54330.
Last modified: 15 Apr 2020 | 5:09:22 UTC

Server is working fine. I'm getting a continuous flow of work. Just make sure your drivers are up to date and compatible (nvidia). Looking at your machines, one has a 2080Ti and had work but didn't start the work in time. The other, AMD machine, doesn't have a GPU.
____________

Chris Raisin
Avatar
Send message
Joined: 14 Apr 20
Posts: 4
Credit: 1,284,935
RAC: 11,723
Level
Ala
Scientific publications
wat
Message 54345 - Posted: 16 Apr 2020 | 5:46:56 UTC
Last modified: 16 Apr 2020 | 5:48:21 UTC

What I am seeing is the following (I belong to Team BOINC Australia but it is not showing).
I have just joined project but nothing is happening although lots of other projects have been running fine!
Please, can you advise what action I should take?

[img][/img]
____________

Chris Raisin
Avatar
Send message
Joined: 14 Apr 20
Posts: 4
Credit: 1,284,935
RAC: 11,723
Level
Ala
Scientific publications
wat
Message 54346 - Posted: 16 Apr 2020 | 5:47:43 UTC - in response to Message 54345.
Last modified: 16 Apr 2020 | 6:18:21 UTC

How do I insert image?

The data showing in my BOINC Manager grid for this Project is:

Project GPUGRID
Account craisin
Team
Work Done 2,359,680,096
Avg. Work Done 30,050,028.77
Resource Share 100 (7.14%)
Status Will remove when tasks done
____________

kain
Send message
Joined: 3 Sep 14
Posts: 152
Credit: 595,591,201
RAC: 662,944
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 54352 - Posted: 16 Apr 2020 | 10:29:55 UTC - in response to Message 54346.

GPUGRID works only on nvidia GPUs, you have AMD.

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 1011
Credit: 2,755,503,447
RAC: 3,180,172
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 54531 - Posted: 2 May 2020 | 13:42:27 UTC

Toni, I think this needs doing again:

02/05/2020 14:40:21 | GPUGRID | [error] Error reported by file upload server: can't write file /home/ps3grid/projects/PS3GRID/upload/321/3nrhA00_450_4-TONI_MDADpr4sn-8-10-RND7238_0_1: No space left on server

Profile ServicEnginIC
Avatar
Send message
Joined: 24 Sep 10
Posts: 231
Credit: 1,549,235,332
RAC: 1,973,881
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 54533 - Posted: 2 May 2020 | 15:08:13 UTC - in response to Message 54531.

Same here for several WUs:

Sat 02 May 2020 15:51:52 WEST | GPUGRID | Started upload of 3i0nA02_450_3-TONI_MDADpr4si-8-10-RND2568_1_10
Sat 02 May 2020 15:51:53 WEST | GPUGRID | [error] Error reported by file upload server: Server is out of disk space
Sat 02 May 2020 15:51:53 WEST | GPUGRID | [error] Error reported by file upload server: can't write file /home/ps3grid/projects/PS3GRID/upload/d7/3i0nA02_450_3-TONI_MDADpr4si-8-10-RND2568_1_10: No space left on server
Sat 02 May 2020 15:51:53 WEST | GPUGRID | Temporarily failed upload of 3i0nA02_450_3-TONI_MDADpr4si-8-10-RND2568_1_9: transient upload error
Sat 02 May 2020 15:51:53 WEST | GPUGRID | Backing off 00:03:37 on upload of 3i0nA02_450_3-TONI_MDADpr4si-8-10-RND2568_1_9

Zirma
Send message
Joined: 21 Apr 20
Posts: 13
Credit: 4,411,884
RAC: 1
Level
Ala
Scientific publications
wat
Message 54535 - Posted: 2 May 2020 | 15:22:35 UTC - in response to Message 54531.

cant send finish work..

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 1011
Credit: 2,755,503,447
RAC: 3,180,172
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 54541 - Posted: 2 May 2020 | 21:41:30 UTC

All uploaded now, but the scheduler is still disabled.

Profile ServicEnginIC
Avatar
Send message
Joined: 24 Sep 10
Posts: 231
Credit: 1,549,235,332
RAC: 1,973,881
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 54542 - Posted: 2 May 2020 | 22:44:25 UTC - in response to Message 54541.

All uploaded now, but the scheduler is still disabled.

Yes, all uploaded tasks keep waiting for scheduler be enabled to be reported as finished.
In the meanwhile, they look at BOINC Manager something like this:



(On triple GPU system #480458)

Clive
Send message
Joined: 2 Jul 19
Posts: 20
Credit: 30,105,415
RAC: 150,616
Level
Val
Scientific publications
wat
Message 54543 - Posted: 3 May 2020 | 3:50:44 UTC

Hi:

On my PC I am limited to 2 GPU WUs. Can this be increased so that there are extra WUs for times like this?

Clive Hunt

Profile ServicEnginIC
Avatar
Send message
Joined: 24 Sep 10
Posts: 231
Credit: 1,549,235,332
RAC: 1,973,881
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 54545 - Posted: 3 May 2020 | 8:51:58 UTC - in response to Message 54543.
Last modified: 3 May 2020 | 8:53:29 UTC

On my PC I am limited to 2 GPU WUs. Can this be increased so that there are extra WUs for times like this?

This is a question that we all GPUGrid's contributors have wondered at any time, but:
Currently, the project's policy is: Two WUs at a time for every single GPU.
This is how it currently looks at my single GPU system #186626:



The server has had to be stopped due to a data overflow.
If policy were (let's suppose) 10 WUs at a time for every GPU, when server were enabled again, it would have 5 times more finished WUs waiting to be validated...
For sure, project managers have pondered about it, and they have probably conluded that it would suppose quickly a new server overflow.
And every single system would take 5 times as long to process that greater number of stored WUs, which could slow down the work.
Many times I think: Life is a compromise. I have alternative projects to process when GPUGrid is stopped...

(PS: If scheduler remains stopped beyond today's 10:49:41 UTC, both finished WUs in above example are loosing their 24H bonus.)

Geralt
Send message
Joined: 14 Feb 16
Posts: 4
Credit: 76,170
RAC: 112
Level

Scientific publications
wat
Message 54546 - Posted: 3 May 2020 | 9:16:01 UTC - in response to Message 54545.

On my PC I am limited to 2 GPU WUs. Can this be increased so that there are extra WUs for times like this?

This is a question that we all GPUGrid's contributors have wondered at any time, but:
Currently, the project's policy is: Two WUs at a time for every single GPU.
This is how it currently looks at my single GPU system #186626:



The server has had to be stopped due to a data overflow.
If policy were (let's suppose) 10 WUs at a time for every GPU, when server were enabled again, it would have 5 times more finished WUs waiting to be validated...
For sure, project managers have pondered about it, and they have probably conluded that it would suppose quickly a new server overflow.
And every single system would take 5 times as long to process that greater number of stored WUs, which could slow down the work.
Many times I think: Life is a compromise. I have alternative projects to process when GPUGrid is stopped...

(PS: If scheduler remains stopped beyond today's 10:49:41 UTC, both finished WUs in above example are loosing their 24H bonus.)



I think this happened with Folding@Home a while back during the covid surge. Their method of distribution was by the use of a collection server and a assignment server. The collection server was down because of an issue (I forgot if it was a disk space issue or a technical issue) but the assignment server kept assigning work units to nodes. When the collection server finally came back up, it was immediately overwhelmed by the amount of work units being uploaded at the same time

djkra
Send message
Joined: 27 Feb 20
Posts: 2
Credit: 10,530,414
RAC: 0
Level
Pro
Scientific publications
wat
Message 54974 - Posted: 27 May 2020 | 0:41:05 UTC - in response to Message 53986.

thankyou

Post to thread

Message boards : News : Server disk full being addressed