How to remove aborted work units?

留言板 : Number crunching : How to remove aborted work units?
留言板合理

To post messages, you must log in.

作者消息
Profile Paul D. Buck
志愿者测试人员

发送消息
已加入:19 Jul 00
贴子:3898
积分:1,158,042
近期平均积分:0
United States
消息 167193 - 发表于:13 Sep 2005, 14:48:46 UTC

One of the drawbacks in using a "package" is that you cannot always control how things are done. In this case, the search function is provided by the MediaWiki package and I have to live with it. The good news is that they are in progress with a new version ...
ID: 167193 · 举报违规帖子
Profile RichaG
志愿者测试人员
Avatar

发送消息
已加入:20 May 99
贴子:1690
积分:19,287,294
近期平均积分:36
United States
消息 166983 - 发表于:13 Sep 2005, 1:00:26 UTC - 回复消息 166536.  

I did it once and posted how to the post requesting it in the science forum but it's better suited for here and it's probably easier to understand here as well.

1. Suspend all projects except the project with the aborted WU in the "Projects" tab.
2. Go to the "work" tab and then manually suspend every WU for the project with the aborted WU.
3. Click on the aborted WU and then on the resume button so that it will attempt to crunch and change from "aborted" to "computational error".
4. Update the project on the "projects" tab and the WU will report and dissappear.
5. Resume all suspended WU's in the "Work" tab and all projects on the "Projects" tab.

This is because BOINC skips over an aborted WU and it will only report if it has been attempted to be crunched.


On step 2, you only need to suspend the oldest workunits of the project that has the abort work units. The project will try to compute the oldest one first and goes through all of the suspended and aborted units unit it gets one that will crunch.

Red Bull Air Racing

Gas price by zip at Seti

ID: 166983 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166982 - 发表于:13 Sep 2005, 0:59:08 UTC - 回复消息 166978.  
最近的修改日期:13 Sep 2005, 1:00:57 UTC

Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

Hmmm, yeah, but where?

Well, for what it is worth, I added it to the work tab area and added the rest of the status values.


While searching the Wiki before posting the question here, I searched on the term "abort".

Eight pages are returned & now I see RDC's procedure listed on the 6th page of the search results.

Scanning the results, #1 and #6 appear to be the most relevant to this issue, so I would read those first:
http://boinc-doc.net/boinc-wiki/index.php?search=abort&go=Go

For some reason, the BOINC FAQ is not returned as one of the eight results, but I can find the procedure listed there by manually browsing to the page.

Thank you, Mr. Buck, for all of your contributions.
This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166982 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166978 - 发表于:13 Sep 2005, 0:51:43 UTC - 回复消息 166803.  
最近的修改日期:13 Sep 2005, 0:58:38 UTC

Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

Hmmm, yeah, but where?

Well, for what it is worth, I added it to the work tab area and added the rest of the status values.


While searching the Wiki before posting the question here, I searched on the term "abort".

Eight pages are returned & now I see RDC's procedure listed on the 6th page of the search results Scanning the reults, #2 and #6 appear to be the most relevant to this issue. http://boinc-doc.net/boinc-wiki/index.php?search=abort&go=Go

For some reason, the BOINC FAQ is not returned as one of the eight results, but I can find the procedure listed there by manually browsing to the page.

Thank you, Mr. Buck, for all of your contributions.
This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166978 · 举报违规帖子
Profile Paul D. Buck
志愿者测试人员

发送消息
已加入:19 Jul 00
贴子:3898
积分:1,158,042
近期平均积分:0
United States
消息 166892 - 发表于:12 Sep 2005, 20:59:32 UTC - 回复消息 166867.  

FAQ -> Questions and Problems seems the right place.

Ok, there too ...
ID: 166892 · 举报违规帖子
Sergey Broudkov
Avatar

发送消息
已加入:24 May 04
贴子:221
积分:561,897
近期平均积分:0
Russia
消息 166867 - 发表于:12 Sep 2005, 19:30:08 UTC - 回复消息 166803.  

Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

Hmmm, yeah, but where?


FAQ -> Questions and Problems seems the right place.
Kitty@SETI team (Russia). Our cats also want to know if there is ETI out there
ID: 166867 · 举报违规帖子
Profile RDC
志愿者测试人员
Avatar

发送消息
已加入:17 May 99
贴子:544
积分:1,215,728
近期平均积分:0
United States
消息 166866 - 发表于:12 Sep 2005, 19:23:55 UTC - 回复消息 166803.  

Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

Hmmm, yeah, but where?

Well, for what it is worth, I added it to the work tab area and added the rest of the status values.


Wow, I feel honored that my solution was good enough to be added to the Wiki. Thanks Paul :)


To truly explore, one must keep an open mind...
ID: 166866 · 举报违规帖子
Profile Paul D. Buck
志愿者测试人员

发送消息
已加入:19 Jul 00
贴子:3898
积分:1,158,042
近期平均积分:0
United States
消息 166803 - 发表于:12 Sep 2005, 15:47:10 UTC - 回复消息 166545.  
最近的修改日期:12 Sep 2005, 16:20:01 UTC

Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

Hmmm, yeah, but where?

Well, for what it is worth, I added it to the work tab area and added the rest of the status values.
ID: 166803 · 举报违规帖子
Profile RDC
志愿者测试人员
Avatar

发送消息
已加入:17 May 99
贴子:544
积分:1,215,728
近期平均积分:0
United States
消息 166553 - 发表于:11 Sep 2005, 23:55:31 UTC

Glad to help out :)


To truly explore, one must keep an open mind...
ID: 166553 · 举报违规帖子
Ken Phillips m0mcw
志愿者测试人员
Avatar

发送消息
已加入:2 Feb 00
贴子:267
积分:415,678
近期平均积分:0
United Kingdom
消息 166546 - 发表于:11 Sep 2005, 23:33:42 UTC - 回复消息 166536.  
最近的修改日期:11 Sep 2005, 23:35:38 UTC

I did it once and posted how to the post requesting it in the science forum but it's better suited for here and it's probably easier to understand here as well.

1. Suspend all projects except the project with the aborted WU in the "Projects" tab.
2. Go to the "work" tab and then manually suspend every WU for the project with the aborted WU.
3. Click on the aborted WU and then on the resume button so that it will attempt to crunch and change from "aborted" to "computational error".
4. Update the project on the "projects" tab and the WU will report and dissappear.
5. Resume all suspended WU's in the "Work" tab and all projects on the "Projects" tab.

This is because BOINC skips over an aborted WU and it will only report if it has been attempted to be crunched.



RDC

You are a star! You have just enabled me to get rid of a troublesome CPDN unit from about two months ago, cheers!

Ken P.

Ken Phillips

BOINC question? Look here



"The beginning is the most important part of the work." - Plato
ID: 166546 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166545 - 发表于:11 Sep 2005, 23:32:43 UTC - 回复消息 166536.  
最近的修改日期:11 Sep 2005, 23:32:58 UTC

I did it once and posted how to the post requesting it in the science forum but it's better suited for here and it's probably easier to understand here as well.

1. Suspend all projects except the project with the aborted WU in the "Projects" tab.
2. Go to the "work" tab and then manually suspend every WU for the project with the aborted WU.
3. Click on the aborted WU and then on the resume button so that it will attempt to crunch and change from "aborted" to "computational error".
4. Update the project on the "projects" tab and the WU will report and dissappear.
5. Resume all suspended WU's in the "Work" tab and all projects on the "Projects" tab.

This is because BOINC skips over an aborted WU and it will only report if it has been attempted to be crunched.



Wow, that's great!
Not exactly intuitive, but it worked.
Sounds like a potential Wiki entry for Mr. Buck.

Thanks for the help!

This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166545 · 举报违规帖子
Profile RDC
志愿者测试人员
Avatar

发送消息
已加入:17 May 99
贴子:544
积分:1,215,728
近期平均积分:0
United States
消息 166536 - 发表于:11 Sep 2005, 23:07:39 UTC

I did it once and posted how to the post requesting it in the science forum but it's better suited for here and it's probably easier to understand here as well.

1. Suspend all projects except the project with the aborted WU in the "Projects" tab.
2. Go to the "work" tab and then manually suspend every WU for the project with the aborted WU.
3. Click on the aborted WU and then on the resume button so that it will attempt to crunch and change from "aborted" to "computational error".
4. Update the project on the "projects" tab and the WU will report and dissappear.
5. Resume all suspended WU's in the "Work" tab and all projects on the "Projects" tab.

This is because BOINC skips over an aborted WU and it will only report if it has been attempted to be crunched.




To truly explore, one must keep an open mind...
ID: 166536 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166531 - 发表于:11 Sep 2005, 22:59:45 UTC
最近的修改日期:11 Sep 2005, 23:01:14 UTC

Also, I should mention the work cache was set to 7 days until this morning.
I set the preferences to communicate every 0.1 days, updated the project to receive the new preferences, aborted the work units, then hit Update a few times to see if that would transfer the aborted work units to the server.
Since that time, I left the Update button alone thinking I might be interfering with the normal transfer schedule.

This all happened because I was trying to get some work units for the faster Windows PC that is out of work. With 2 PC's, one fast and one slower, and the work cache set to a high number, the fast PC always seems to run out of work & then wait for the slow PC to finish & upload all of its results before trying to get any more from the server. I figured aborting the work units would throw them back on the heap for somebody else to fill the quorum.

With the fixes put in place by Matt L. & Co., I'm trying to get back to a smaller work cache.
This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166531 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166523 - 发表于:11 Sep 2005, 22:48:04 UTC - 回复消息 166517.  
最近的修改日期:11 Sep 2005, 22:50:53 UTC

have the files uploaded or are they still listed on the "transfers" tab? Are the scheduler RPCs actually getting through? What does the messages tab have to say?


No transfers waiting to go for the aborted WU's. They don't seem to be generating any transfers for the scheduler to pick up. The last message for each of the aborted ones was Unrecoverable error for work unit (Aborted by GUI).

Since that time, the one WU that was crunching has finished and the only one waiting to run has started and finished.

By the way, this is happening with the optimized client for linux but the same thing is happening with an aborted CPDN WU on a windows PC. The aborted WU's just don't tell the server they are aborted.

Hitting Update communicates successfully with the scheduler (0 units requested, 0 received).
This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166523 · 举报违规帖子
Profile Toby
志愿者测试人员
Avatar

发送消息
已加入:26 Oct 00
贴子:1005
积分:6,366,949
近期平均积分:0
United States
消息 166517 - 发表于:11 Sep 2005, 22:37:03 UTC

have the files uploaded or are they still listed on the "transfers" tab? Are the scheduler RPCs actually getting through? What does the messages tab have to say?
A member of The Knights Who Say NI!
For rankings, history graphs and more, check out:
My BOINC stats site
ID: 166517 · 举报违规帖子
Colossus
Avatar

发送消息
已加入:8 Jul 05
贴子:73
积分:23,618
近期平均积分:0
United States
消息 166513 - 发表于:11 Sep 2005, 22:20:32 UTC

Anybody know the best way to delete work units that have been aborted but won't upload to the scheduler?

They've been sitting there all day with the work cache set to connect every 0.1 days. Hitting Update a few times in the first hour didn't seem to help.
This is the voice of world control. I bring you peace. It may be the peace of plenty and content or the peace of unburied dead. The choice is yours.
ID: 166513 · 举报违规帖子

留言板 : Number crunching : How to remove aborted work units?


 
©2020 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.