| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: I've been 'Detached'! |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I had to revert to a back-up copy of my boinc installation recently, and then subsequently managed to retrieve a more recent copy.
Yesterday, when reporting the work done, all the work units that were reported from the later copy (and I mean spent an age uploading and succesfully reported) have been marked as 'detached' instead. (Note, as far as the newer installation was concerned, these work units were 'Ready to Run') Is this something I did and can I avoid it in the future - assuming this happens again! |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Yes you did it. 2 possibles I can think of:
----------------------------------------1 you installed a second copy, or at least created a different data_directory, so one client told the servers you no longer intended to do the work and the other copy still had the work, completed it and send it up. OR, 2 you installed new client, loosing the work, then restored and started on the lost work. Anyhow at some point in time the server was informed of a detach event and resubmitted the work to other volunteers. In the cause of this you may have created a new device on your Device Manager list.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Definitely didn't do the second.
For the first option, do you mean the 'BOINC' folder having a different name? That might be possible as I (obviously) had two copies at one point... Or would a different path be enough to cause this problem? Thanks for the help, MacDitch |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
A different name/location/drive.
----------------------------------------The BOINC instraller is supposed to check if an old data dir exists, but BOINC being BOINC and changes made at will, me myself have experienced more than one install where BOINC forgot the data, so I migrated/moved the data myself. AND, religiously I do this off-line, so the client does not have a chance to talk to the servers. If the new client/version suddenly forgot, at least this way the victim can move the data_dir and all it's content (while the client is of course fully exited, no service running).
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: 
|
I had to revert to a back-up copy of my boinc installation recently, and then subsequently managed to retrieve a more recent copy. So, just to make sure understood you correctly, let's example say you: 1; Restored a 2-day old BOINC-client. 2; Aferwards restored a 1-day old BOINC-client. Now, to BOINC it doesn't matter, if either #1 or #2 connected to projects scheduling-server, all work currently assigned to you would have been marked "client detached". Since this is marked server-side, it wouldn't matter if you restored a later copy and in client the work was marked "Ready to Run", it's still marked "detached" on server. Since WCG does run with server-aborts turned-on, any client-connection to scheduling-server should abort all detached work on client, as long as you're using v5.8.17 or later. BTW, how old the restored BOINC-copy is doesn't really matter, since if you do: a; Makes a backup-copy. b; Does a scheduler-rpc. c; Restores backup-copy made 1 minute ago. You'll get "client detached" next time the restored backup-copy makes a scheduler-connection.  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
ergo, to make a long story short, do as I do this, whilst off-line and the contact counters won't increment or revert.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
ergo, to make a long story short, do as I do this, whilst off-line and the contact counters won't increment or revert. Well, you normally don't know beforehand you'll need to restore a backup-copy or not... In WCG there most wu's takes less than 12 hours to run, in most instances a backup-copy of BOINC is worthless, since chances are whatever you've backed-up is already reported, or isn't started yet and can be re-issued by server. "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So, just to make sure understood you correctly, let's example say you: 1; Restored a 2-day old BOINC-client. 2; Aferwards restored a 1-day old BOINC-client. Exactly. In WCG there most wu's takes less than 12 hours to run, in most instances a backup-copy of BOINC is worthless, since chances are whatever you've backed-up is already reported, or isn't started yet and can be re-issued by server. In this case the computer only connects once a week and had completed about four days before the 'hiccup'. End result was that I had two 'sets' of results to upload (I didn't know I could def. get the older copy initially and therefore worked on the later copy) and the problem was when I tried to report the second batch. I'll notch it up as a lesson learnt and be more aware in the future. |
||
|
|
|