- Joined
- Jan 1, 2005
- Messages
- 5,459
- Reaction score
- 21
Thre is a problem with certain Genome Comparison units that get stuck at 0% for ever until aborted.
(I discovered one on one of my machines this morning that had crunched for over 24hrs without progress )
The WCG Techs have issued a server code that will automatically abort the troublesome units if using Boinc 5.8.xx - if using earlier versions of Boinc the offending unit numbers will be notified in your messages tab, but you will have to manually select and abort them.
I urge all teamies to do a manual update of their Boinc clients and check their machines
(I discovered one on one of my machines this morning that had crunched for over 24hrs without progress )
The WCG Techs have issued a server code that will automatically abort the troublesome units if using Boinc 5.8.xx - if using earlier versions of Boinc the offending unit numbers will be notified in your messages tab, but you will have to manually select and abort them.
The BOINC 5.8 client introduced a new feature where the BOINC server can send a message to it telling it to abort a workunit. However, there wasn't any server side code to take advantage of the this new message. We talked to the BOINC developers over the past day and we have implemented a new feature on the server that allows the server to abort a workunit on a client if the workunit has been canceled on the server. We deployed this to our servers about 3 hours ago and we have just finished canceling all of the suspect workunits.
This means that any member using the BOINC client 5.8 will get a message telling it to abort any workunits that are part of this problem the next time that the client contacts the server. Go ahead and manually force an update to have your client contact the server immediately by going to the advanced view, project tab and selecting World Community Grid and then click on 'Update'.
If you have one of the troublesome workunits you will see a message such as:
3/21/2007 4:43:16 PM|World Community Grid|Message from server: Result 10000001-10001441_0 is no longer usable
and you will see in the tasks tab that the result is now marked 'Aborted by project'.
I should also mention that earlier versions of the BOINC client will still see the message:
3/21/2007 4:43:16 PM|World Community Grid|Message from server: Result 10000001-10001441_0 is no longer usable
but the automatic abort will not occur. Users who see this messages should aborted the referenced workunit.
I urge all teamies to do a manual update of their Boinc clients and check their machines
Last edited: