lots of ghost WU

Message boards : Number crunching : lots of ghost WU

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Christian Hagen

Send message
Joined: 26 Sep 05
Posts: 5
Credit: 46,795
RAC: 0
Message 980 - Posted: 5 Oct 2005, 13:44:58 UTC


hello everybody.

i just had a look at my results. according to that page i have about 100 WU's sent to my computer. but actually there are only about 6 here. i conclude that the other WU's are socalled ghosts (already appeared in other BOINC project, e.g., S@h).

So my questions are:

- Does somebody else have similar experiences?

- Is there a possibility to avoid such ghosts?
ID: 980 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile FZB

Send message
Joined: 17 Sep 05
Posts: 84
Credit: 4,948,999
RAC: 0
Message 984 - Posted: 5 Oct 2005, 15:45:51 UTC
Last modified: 5 Oct 2005, 15:52:10 UTC

no ghost units for me so far (in any project, though i think there was a issue with a CC version as far as i remember), what boinc CC version are you using? manager -> help -> about
--
Florian
www.domplatz1.de
ID: 984 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Christian Hagen

Send message
Joined: 26 Sep 05
Posts: 5
Credit: 46,795
RAC: 0
Message 986 - Posted: 5 Oct 2005, 15:57:08 UTC - in response to Message 984.  

no ghost units for me so far (in any project, though i think there was a issue with a CC version as far as i remember), what boinc CC version are you using? manager -> help -> about


i'm using version 4.43 at the moment.
ID: 986 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile FZB

Send message
Joined: 17 Sep 05
Posts: 84
Credit: 4,948,999
RAC: 0
Message 1017 - Posted: 6 Oct 2005, 7:33:08 UTC

what os are you using? if windows, the recommendend version would be 4.45. on the burp forums i read 4.4X has the ghost problem, if you want to try a beta version (runs stable here and have not read too much trouble on the forums), you might want to try 4.72
win: http://boinc.berkeley.edu/dl/boinc_4.72_windows_intelx86.exe
linux: http://boinc.berkeley.edu/dl/boinc_4.72_i686-pc-linux-gnu.sh
mac: not sure what the latest "stable" beta is there

on the other hand, with a recommended 5.X version coming soon, you might wait for that if you are not comfortable with beta software.
--
Florian
www.domplatz1.de
ID: 1017 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Christian Hagen

Send message
Joined: 26 Sep 05
Posts: 5
Credit: 46,795
RAC: 0
Message 1053 - Posted: 7 Oct 2005, 7:09:47 UTC - in response to Message 1017.  

what os are you using? if windows, the recommendend version would be 4.45. on the burp forums i read 4.4X has the ghost problem, if you want to try a beta version (runs stable here and have not read too much trouble on the forums), you might want to try 4.72
win: http://boinc.berkeley.edu/dl/boinc_4.72_windows_intelx86.exe
linux: http://boinc.berkeley.edu/dl/boinc_4.72_i686-pc-linux-gnu.sh
mac: not sure what the latest "stable" beta is there

on the other hand, with a recommended 5.X version coming soon, you might wait for that if you are not comfortable with beta software.


Thanks for the help, but i think i'll skip the beta versions and wait for version 5.
ID: 1053 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Robert Nelson

Send message
Joined: 17 Sep 05
Posts: 4
Credit: 1,312,027
RAC: 1,752
Message 1069 - Posted: 7 Oct 2005, 17:09:39 UTC

Version 4.45 solved the ghost problem. 4.43 seemed to have been especially suseptable. Ran 4.45 for quite a while with no problems. Now running both 4.72 and 5.X versions with no sign of that problem either. If you don't wish to use the beta version 4.45 is the standard version and should clear your problem.
ID: 1069 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Divide Overflow

Send message
Joined: 17 Sep 05
Posts: 82
Credit: 921,382
RAC: 0
Message 1076 - Posted: 7 Oct 2005, 20:10:45 UTC - in response to Message 1069.  
Last modified: 7 Oct 2005, 20:13:35 UTC

If you don't wish to use the beta version 4.45 is the standard version and should clear your problem.

Not if you're using Linux (as Christian is). v4.43 is the current recommended version for Linux users.
ID: 1076 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Fuzzy Hollynoodles
Avatar

Send message
Joined: 7 Oct 05
Posts: 234
Credit: 15,020
RAC: 0
Message 1091 - Posted: 8 Oct 2005, 2:19:27 UTC

I have the 4.72 version and I'm very satisfied with it! It's very stable and has been since I installed it.'

I have heard there was a lot of trouble with the 5.* client, and since my pet project LHC can't run on anything higher than 4.*, I'm sticking to my 4.72 client.
[b]"I'm trying to maintain a shred of dignity in this world." - Me[/b]

ID: 1091 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Paul D. Buck

Send message
Joined: 17 Sep 05
Posts: 815
Credit: 1,812,737
RAC: 0
Message 1093 - Posted: 8 Oct 2005, 5:45:34 UTC

for the moment, me too ... :)
ID: 1093 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Rebirther
Avatar

Send message
Joined: 17 Sep 05
Posts: 116
Credit: 41,315
RAC: 0
Message 1102 - Posted: 8 Oct 2005, 10:05:55 UTC

After trying download new work with a server error -500 got with the next successful download 7 ghost WUs with Boinc 5.1.8.
ID: 1102 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Robert Nelson

Send message
Joined: 17 Sep 05
Posts: 4
Credit: 1,312,027
RAC: 1,752
Message 1104 - Posted: 8 Oct 2005, 12:41:51 UTC - in response to Message 1076.  
Last modified: 8 Oct 2005, 12:45:04 UTC

If you don't wish to use the beta version 4.45 is the standard version and should clear your problem.

Not if you're using Linux (as Christian is). v4.43 is the current recommended version for Linux users.
Thanks David, thats too bad they didn't get around to putting that version for Linux, the ghostbuster fix was extremely helpful.
After trying download new work with a server error -500 got with the next successful download 7 ghost WUs with Boinc 5.1.8.

Argghhh had not got around to trying 5.1.8, the new versions were coming out way to fast, but if they broke the fix in that one, its time to call Ghost busters. Will wait for 5.1.9 to show up.
I have the 4.72 version and I'm very satisfied with it! It's very stable and has been since I installed it.'

I have heard there was a lot of trouble with the 5.* client, and since my pet project LHC can't run on anything higher than 4.*, I'm sticking to my 4.72 client.

Agree on 4.72. I am running that on some boxes as well, so that I can continue to crunch LHC and Sztaki. Have yet to see anything on the 5 versions that caused alarm, except the add project wizard, which I don't have to use very often.



ID: 1104 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Rebirther
Avatar

Send message
Joined: 17 Sep 05
Posts: 116
Credit: 41,315
RAC: 0
Message 1568 - Posted: 21 Oct 2005, 14:08:34 UTC

Here again, got a server error -500 by trying to get new work with Boinc 5.2.2, not one WU arrived but in list marked "sent out" :(
ID: 1568 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Robert Nelson

Send message
Joined: 17 Sep 05
Posts: 4
Credit: 1,312,027
RAC: 1,752
Message 1621 - Posted: 22 Oct 2005, 14:50:28 UTC - in response to Message 1568.  

Here again, got a server error -500 by trying to get new work with Boinc 5.2.2, not one WU arrived but in list marked "sent out" :(

Checked my accounts at Rosetta and other projects running. It appears that I may have received two ghost work units on Oct 14 from Rosetta. Looking at other projects don't seem to find any evidence of any on the other projects. The ghost units were on a Windows computer running BOINC version 5.1.4 at the time, the Rosetta units were version 4.77. Broadband connection to internet. My previous experience with the ghost was that these type of beasties appeared in batches. So either another fix broke it again, or we have another species out there. Now running 5.2.2 on all but one machine, will watch for the poltergeist closely.
ID: 1621 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Rebirther
Avatar

Send message
Joined: 17 Sep 05
Posts: 116
Credit: 41,315
RAC: 0
Message 4660 - Posted: 29 Nov 2005, 8:22:36 UTC

The problem is still running by me:
29/11/2005 09:15:07|rosetta@home|Sending scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi
29/11/2005 09:15:07|rosetta@home|Reason: To fetch work
29/11/2005 09:15:07|rosetta@home|Requesting 17251 seconds of new work
29/11/2005 09:15:12|rosetta@home|Scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi failed with a return value of 500
29/11/2005 09:15:12|rosetta@home|No schedulers responded

After that WUs are listed as sent out but nothing received, the second time Iam getting a "no work from project" message but then getting new work, its only by the rosetta project all other projects are working fine.
ID: 4660 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Nothing But Idle Time

Send message
Joined: 28 Sep 05
Posts: 209
Credit: 139,545
RAC: 0
Message 5108 - Posted: 4 Dec 2005, 15:14:08 UTC - in response to Message 4660.  
Last modified: 4 Dec 2005, 15:17:43 UTC

The problem is still running by me:
29/11/2005 09:15:07|rosetta@home|Sending scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi
29/11/2005 09:15:07|rosetta@home|Reason: To fetch work
29/11/2005 09:15:07|rosetta@home|Requesting 17251 seconds of new work
29/11/2005 09:15:12|rosetta@home|Scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi failed with a return value of 500
29/11/2005 09:15:12|rosetta@home|No schedulers responded

After that WUs are listed as sent out but nothing received, the second time Iam getting a "no work from project" message but then getting new work, its only by the rosetta project all other projects are working fine.


=====================
I've experienced this same error message for quite some time. I mentioned it in another thread long ago and was told by David Kim that he hadn't heard of anyone reporting this particular error. He suggested I upgrade to the latest v5. Well I've upgraded routinely 5.2.2 to 5.2.13 and still have the same problem. Paul D. Buck suggested it might be my cable modem -- but can't prove that. I now have 4 WUs at LHC (that haven't expired and been deleted) and have 17 Wus here at Rosetta that I was supposedly sent but I know I never received. At times like this I wish I had a "Request Resend" button so I could dispose of these WUs. My download error rate seems proportional to the number of times/day that I request new work, yet I never get this particular problem with Einstein@Home (why not?). If this is BOINC software related I wish someone would fix it.


ID: 5108 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Golden Turtle

Send message
Joined: 23 Sep 05
Posts: 34
Credit: 22,941
RAC: 0
Message 5110 - Posted: 4 Dec 2005, 15:42:18 UTC

I had a lot of 'Ghosts' reported in -Questions and Problems- Nov 16 & 17 msge 3440 etc but have received no response from admin, am now running the latest 5... and the problem has probably gone away but those WUs with around Dec 15 deadlines still show in my account as outstanding.
ID: 5110 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Nothing But Idle Time

Send message
Joined: 28 Sep 05
Posts: 209
Credit: 139,545
RAC: 0
Message 5123 - Posted: 4 Dec 2005, 18:44:25 UTC
Last modified: 4 Dec 2005, 18:44:56 UTC

Not news to others I suppose, but I just discovered that Einstein@Home has software that resyncs the server and client so that "ghosts" or otherwise missing WUs are resent. Refer to thread.

I guess Rosetta and LHC don't think that resending missing Wus is necessary; not for me to judge. But I personally am more than a little annoyed at these missing Wus because I feel it reflects on me, my setup, and my ability to produce an output.
ID: 5123 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Rebirther
Avatar

Send message
Joined: 17 Sep 05
Posts: 116
Credit: 41,315
RAC: 0
Message 5126 - Posted: 4 Dec 2005, 19:23:02 UTC

Iam using now Boinc 5.2.13 and no problems so far with these ghost WUs and error -500 since some days.
ID: 5126 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Tern
Avatar

Send message
Joined: 25 Oct 05
Posts: 576
Credit: 4,695,450
RAC: 13
Message 5138 - Posted: 5 Dec 2005, 1:19:08 UTC - in response to Message 5123.  

Einstein@Home has software that resyncs... (snip)

I guess Rosetta and LHC don't think that resending missing Wus is necessary; not for me to judge. But I personally am more than a little annoyed at these missing Wus because I feel it reflects on me, my setup, and my ability to produce an output.


It's a question of whether the additional load on the server to deal with the "resync" is available at all, and if it is, if it's worth it compared to the load caused by resending the result to another host. Einstein has "over-powerful" servers with a lot of capacity, while SETI barely scrapes through - there is no way the SETI servers, which can't even handle calculating "pending credit", could handle resyncing. I don't know where Rosetta falls in this spectrum.

Regardless, network problems happen. They shouldn't reflect on you; I would only worry about "my ability to produce" if the results I were returning were full of errors. THEN I would try to solve the problem on my end, if any. "Download" issues... eh, it could be anything.

ID: 5138 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Paul D. Buck

Send message
Joined: 17 Sep 05
Posts: 815
Credit: 1,812,737
RAC: 0
Message 5157 - Posted: 5 Dec 2005, 8:14:17 UTC

There is a "general" problem occurring with some network activities. This was in the Developers thread. All I can say is some of the best networking minds I know of in the BOINC Community are looking at some of these problems. You just may be one of the unlucky few that have these problems.

The best news is that they are looking for people that are interested in trying to track down the cause of the problem. The two that I know that are involved are Celtic Wolf and Ian Tighe.

So, if you are interested in trying to help locate this problem, will run tests for them, and send them some log information ... well, the world needs YOU!

So, are you interested?
ID: 5157 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
1 · 2 · Next

Message boards : Number crunching : lots of ghost WU



©2024 University of Washington
https://www.bakerlab.org