Rosetta Mimi 1.03 locking up Boinc on windows

Message boards : Current tests : Rosetta Mimi 1.03 locking up Boinc on windows

To post messages, you must log in.

AuthorMessage
[B^S] sTrey
Avatar

Send message
Joined: 15 Feb 06
Posts: 58
Credit: 15,430
RAC: 0
Message 3601 - Posted: 13 Jan 2008, 7:56:20 UTC

What's with the Rosetta Mini 1.03 app? Crashed once, made boinc and boincmgr nonresponsive. Boinc isn't a service on this box so I logged out/in to restart. It restarted this task with high priority, according to the client, but nothing was running and boinc was again stalled.
I had to edit client_state.xml to force the project suspended to get the client running again.

This happened on 2 different pc's before I caught it. The cash dialogue comes up about 19 seconds into the wu.
Clients are 5.10.20
ID: 3601 · Report as offensive    Reply Quote
Profile [BOINC@Poland]emik

Send message
Joined: 4 Jan 07
Posts: 2
Credit: 33,039
RAC: 0
Message 3602 - Posted: 13 Jan 2008, 9:55:30 UTC

me too
ID: 3602 · Report as offensive    Reply Quote
John Hunt
Avatar

Send message
Joined: 16 Mar 07
Posts: 10
Credit: 28,654
RAC: 0
Message 3603 - Posted: 13 Jan 2008, 13:56:43 UTC

Caught me on BOINC 5.10.30 too!
ID: 3603 · Report as offensive    Reply Quote
[B^S] sTrey
Avatar

Send message
Joined: 15 Feb 06
Posts: 58
Credit: 15,430
RAC: 0
Message 3604 - Posted: 13 Jan 2008, 16:51:41 UTC
Last modified: 13 Jan 2008, 17:16:05 UTC

On the first pc mentioned before, BOINC stopped responding sometime overnight and I had to kill it and the boincmgr. Now boinc won't come up. The process appears but the manager can't connect. Hope that doesn't happen to others.

-- Edit: just happened to the 2nd pc also. Must leave the house for much of the day so I don't know what's going on, client_state.xml looks fine, tried rebooting and reinstalling on one of the pc's, this is not a happy crunching day.
ID: 3604 · Report as offensive    Reply Quote
TheGummer

Send message
Joined: 17 Feb 06
Posts: 1
Credit: 11,229
RAC: 0
Message 3605 - Posted: 13 Jan 2008, 18:30:25 UTC


Its also halting version 5.10.35. I've suspended it for now.
ID: 3605 · Report as offensive    Reply Quote
[B^S] sTrey
Avatar

Send message
Joined: 15 Feb 06
Posts: 58
Credit: 15,430
RAC: 0
Message 3606 - Posted: 14 Jan 2008, 2:24:04 UTC

Finally cleaned up after this mess. Make sure nothing is left from the mini app in one of your slots dirs, or you will have a broken boinc (possibly much later) in your future.

Please let us know when we can "go back in the water" without these wus showing up. Thanks



ID: 3606 · Report as offensive    Reply Quote
Dr Who Fan
Avatar

Send message
Joined: 2 Sep 06
Posts: 76
Credit: 107,857
RAC: 0
Message 3611 - Posted: 14 Jan 2008, 8:21:20 UTC
Last modified: 14 Jan 2008, 8:22:41 UTC

Someone has posted at RALPH@home bug list / Rosetta Min 1.03 basically the same experience.


ID: 3611 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 3617 - Posted: 15 Jan 2008, 15:56:58 UTC

It should be enough to remove the "..Boincslots2minirosetta_database" tree, everything should be fine afterwards. (OK, if there are no uncrunched rosetta_minis in your Boinc tree anymore :-)

Peter
ID: 3617 · Report as offensive    Reply Quote
Profile dekim
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 20 Jan 06
Posts: 250
Credit: 543,579
RAC: 0
Message 3619 - Posted: 15 Jan 2008, 17:01:27 UTC

We are looking into this bug. We will have to send out more work units to help debug. Sorry for the troubles.
ID: 3619 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 3625 - Posted: 15 Jan 2008, 22:56:56 UTC

The saga continues in RALPH@home bug list / Rosetta Min 1.03 thread...

Peter
ID: 3625 · Report as offensive    Reply Quote
Profile WyerByter
Avatar

Send message
Joined: 12 Aug 06
Posts: 1
Credit: 29,790
RAC: 0
Message 3628 - Posted: 16 Jan 2008, 3:02:36 UTC

There was a message on the BOINC Alpha mailing list. It appears that this is due to a strange interaction with Mini and BOINC. Particularly, the previously mentioned tree has some files that are read only (SVN files). When BOINC fails to delete them, it assumes that it has been locked by another process (One that will release shortly) and tries to delete it again. They have put in a fix for Version 6, and Dr Anderson has requested that it be ported back to 5.10.x.

Of course if it was possible to remove those files from the archive, or remove the read only attribute, that would also solve the problem.

I currently have a mini working, I have removed the read only attribute from the whole tree, I will try and post again when it finishes, as to success or failure.
This signiture stolen from somewhere.

ID: 3628 · Report as offensive    Reply Quote
Viking69

Send message
Joined: 21 Feb 06
Posts: 8
Credit: 128,054
RAC: 0
Message 3633 - Posted: 16 Jan 2008, 10:41:06 UTC - in response to Message 3619.  
Last modified: 16 Jan 2008, 10:46:03 UTC

We are looking into this bug. We will have to send out more work units to help debug. Sorry for the troubles.


Don't SEND OUT more WU's of this. It locks up the system. Rosetta and RALPH are bothe suspended for now until I see that this is fixed in this forum and the BOINCalpha forums.

Fix this internally FIRST.

EDIT: never mind. Ii see that in another thread that David A is aware and has a fix. But WHY were these files included in the system anyway?
ID: 3633 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 3634 - Posted: 16 Jan 2008, 11:08:03 UTC - in response to Message 3633.  

Don't SEND OUT more WU's of this. It locks up the system.

Does it really? How exactly? If, then something else must bee happening too. Can you please explain? Noone except me mentioned such system behavior.

Ii see that in another thread that David A is aware and has a fix. But WHY were these files included in the system anyway?

You have read it yourself: by accident.

Peter
ID: 3634 · Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 8 Aug 06
Posts: 70
Credit: 2,279,694
RAC: 0
Message 3640 - Posted: 17 Jan 2008, 3:22:54 UTC

One thing is not clear to me...

Is this really a windows-only problem? If we are talking about read-only files unable to be deleted, doesn't that hit all OS?
Dublin, CA
SETI.USA - Stats - My stuff - BOINC IRC chat
ID: 3640 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 3641 - Posted: 17 Jan 2008, 8:28:45 UTC - in response to Message 3640.  

Is this really a windows-only problem? If we are talking about read-only files unable to be deleted, doesn't that hit all OS?

I think it should pertain to all Boinc versions. Except it did not :-)

What does your Mac say to Rosetta mini?

Possibly the reason is a rather small count of non-Windows hosts to receive AND notice such problematic WU?

Peter
ID: 3641 · Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 8 Aug 06
Posts: 70
Credit: 2,279,694
RAC: 0
Message 3643 - Posted: 17 Jan 2008, 8:46:45 UTC - in response to Message 3641.  

What does your Mac say to Rosetta mini?


Nothing yet. It has not yet started.

Dublin, CA
SETI.USA - Stats - My stuff - BOINC IRC chat
ID: 3643 · Report as offensive    Reply Quote
Jim_

Send message
Joined: 7 Apr 06
Posts: 1
Credit: 10,282
RAC: 0
Message 3644 - Posted: 17 Jan 2008, 17:04:48 UTC

Installed BOINC 5.10.38 on a system (Windows XP) that has been seeing the Rosetta mini "lock-up". The work unit that was running did fail at completion but BOINC didn't exhibit the previously seen problems.

The specific error with Rosetta mini was:
1/17/2008 11:17:46 AM|ralph@home|Computation for task mini_abrelax-1scjB-test_james_2847_3_1 finished
1/17/2008 11:17:46 AM|ralph@home|Output file mini_abrelax-1scjB-test_james_2847_3_1_0 for task mini_abrelax-1scjB-test_james_2847_3_1 absent

Note that I had not changed attributes on any files in the slots directory.

I don't have any Rosetta mini work queued and Ralph currently doesn't have any work for this system so I can't tell if the problem seen will reoccur or if it was an outgrowth of all the terminating of BOINC I forced to do when BOINC would stall.
ID: 3644 · Report as offensive    Reply Quote
Billy

Send message
Joined: 29 Jan 07
Posts: 14
Credit: 7,865
RAC: 0
Message 3651 - Posted: 20 Jan 2008, 7:00:42 UTC

Just had one lock up on my Intel Core Duo Mac running OSX 10.4.11 and Boinc 5.10.34
ID: 3651 · Report as offensive    Reply Quote

Message boards : Current tests : Rosetta Mimi 1.03 locking up Boinc on windows



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