I was finally able to upload the workunit my computer finished yesterday, but not the one it finished today. Attempting to upload it gave some unfamiliar messages:

12/1/2008 11:43:36 PM|rosetta@home|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 1 completed tasks
12/1/2008 11:44:07 PM|rosetta@home|Scheduler request succeeded: got 0 new tasks
12/1/2008 11:44:07 PM|rosetta@home|Message from server: Incomplete request received.
12/1/2008 11:44:07 PM|rosetta@home|New host venue:

For those that are still having issues with your boinc client not getting the new scheduler url from our master url and if you would rather not detach and reattach the project which should fix the issue if updating doesn't. You can do the following:

1. stop the boinc client.
2. manually edit the client_state.xml, client_state_prev.xml, and master_boinc.bakerlab.org_rosetta.xml files in the BOINC client data directory by changing all occurrences of "https://boinc.bakerlab.org/rosetta_cgi/cgi" to "http://srv4.bakerlab.org/rosetta_cgi/cgi". This should be between and tags. You might not need to update master_boinc.bakerlab.org_rosetta.xml but you might as well.

If anyone has any suggestions or ideas for an easier fix, I'm all ears. The redirection does not appear to work so I removed it.

For those that are still having issues ...

I can confirm this worked for me. Thanks!

For those that are still having issues with your boinc client not getting the new scheduler url from our master url and if you would rather not detach and reattach the project which should fix the issue if updating doesn't. You can do the following:

1. stop the boinc client.
2. manually edit the client_state.xml, client_state_prev.xml, and master_boinc.bakerlab.org_rosetta.xml files in the BOINC client data directory by changing all occurrences of "https://boinc.bakerlab.org/rosetta_cgi/cgi" to "http://srv4.bakerlab.org/rosetta_cgi/cgi". This should be between <scheduler> and </scheduler> tags. You might not need to update master_boinc.bakerlab.org_rosetta.xml but you might as well.

If anyone has any suggestions or ideas for an easier fix, I'm all ears. The redirection does not appear to work so I removed it.

Does these suggestions lose all existing work? If so I really don't appreciate losing 4+ cpu hours per WU! Note this would be worst when you change the default to 6 hours.

For those that are still having issues with your boinc client not getting the new scheduler url from our master url and if you would rather not detach and reattach the project which should fix the issue if updating doesn't. You can do the following:

1. stop the boinc client.
2. manually edit the client_state.xml, client_state_prev.xml, and master_boinc.bakerlab.org_rosetta.xml files in the BOINC client data directory by changing all occurrences of "https://boinc.bakerlab.org/rosetta_cgi/cgi" to "http://srv4.bakerlab.org/rosetta_cgi/cgi". This should be between <scheduler> and </scheduler> tags. You might not need to update master_boinc.bakerlab.org_rosetta.xml but you might as well.

If anyone has any suggestions or ideas for an easier fix, I'm all ears. The redirection does not appear to work so I removed it.

Does these suggestions lose all existing work? If so I really don't appreciate losing 4+ cpu hours per WU! Note this would be worst when you change the default to 6 hours.

it is only redirecting your communications/scheduler to the new correct server.
should have no impact on your tasks in process.
only do this if you are still having communications trouble.

The only suggestion I've seen that would cause you to lose any work is if you reset the project. It will lose the work, and start from scratch. Getting the new scheduler URL, and also downloading all of the base files Rosetta uses, for each application version you receive tasks for.

The client will retry, and wait and retry, and eventually double check the master file. At that point it will begin to use the new scheduler URL all by itself. No manual changes required.

---

Rosetta Moderator: Mod.Sense

The only suggestion I've seen that would cause you to lose any work is if you reset the project. It will lose the work, and start from scratch. Getting the new scheduler URL, and also downloading all of the base files Rosetta uses, for each application version you receive tasks for.

The client will retry, and wait and retry, and eventually double check the master file. At that point it will begin to use the new scheduler URL all by itself. No manual changes required.

maybe you guys could put this out on the "news" section so people know.

Hello all,

Came home from work and started my computer.
Boinc replied:
2-12-2008 18:08:09|rosetta@home|Sending scheduler request: To fetch work. Requesting 8884 seconds of work, reporting 0 completed tasks
2-12-2008 18:08:12|rosetta@home|Scheduler request succeeded: got 0 new tasks
2-12-2008 18:08:12|rosetta@home|Message from server: Server error: can't attach shared memory
2-12-2008 18:09:12|rosetta@home|Fetching scheduler list
2-12-2008 18:09:17|rosetta@home|Master file download succeeded
2-12-2008 18:09:22|rosetta@home|Sending scheduler request: To fetch work. Requesting 8546 seconds of work, reporting 0 completed tasks
2-12-2008 18:09:27|rosetta@home|Scheduler request succeeded: got 1 new tasks
2-12-2008 18:09:27|rosetta@home|New host venue: home
2-12-2008 18:09:29|rosetta@home|Started download of boinc_yebf_aah014_03_05.200_v1_3.gz

(Boinc 5.10.45, local time is UTC +1 hour)

Problem solved.

Good luck downloading Wu's,
Path7.

Does these suggestions lose all existing work? If so I really don't appreciate losing 4+ cpu hours per WU!

I can confirm no work is aborted or rejected. 9 WUs successfully UL'd and credited here after making this change.

---

For those that are still having issues with your boinc client not getting the new scheduler url from our master url and if you would rather not detach and reattach the project which should fix the issue if updating doesn't. You can do the following:

1. stop the boinc client.
2. manually edit the client_state.xml, client_state_prev.xml, and master_boinc.bakerlab.org_rosetta.xml files in the BOINC client data directory by changing all occurrences of "https://boinc.bakerlab.org/rosetta_cgi/cgi" to "http://srv4.bakerlab.org/rosetta_cgi/cgi". This should be between <scheduler> and </scheduler> tags. You might not need to update master_boinc.bakerlab.org_rosetta.xml but you might as well.

If anyone has any suggestions or ideas for an easier fix, I'm all ears. The redirection does not appear to work so I removed it.

Have your DNS guy create a CNAME alias for schedular.bakaerlab.org and point that to the A record of whatever machine will be doing the scheduling.

i.e.

$origin bakerlab.org.
schedular CNAME srv4.bakerlab.org.

(don't forget the trailing period)

Then publish the permanent URL for the schedular as http://schedular.bakerlab.org/rosetta_cgi/cgi

Then if you ever need to change the machine that runs the schedular, it's only a global DNS change and does not impact any client... If you set the refresh and timeouts short on the zone these changes can be made to propagate really quickly too... sometimes in a couple of minutes..

---

Looking for a team ??? Join BoincSynergy!!

message from server:progect incountered internal error:shared memory... When I try to update so I can recieve more tasks and have finnished tasks retrieved, I get this message. Is there a problem with rosetta or my wu?

Hi;
I get the same thing... Whenever my computer tries to upload any completed tasks, it receives the message: Message from server: Server error: can't attach shared memory.

I wonder if I did something? But I don't think so, as I have not changed anything in my prefs for either boinc manager or the web site.

Fix it Folks!

Larry

---

LarryB56

message from server:progect incountered internal error:shared memory... When I try to update so I can recieve more tasks and have finnished tasks retrieved, I get this message. Is there a problem with rosetta or my wu?

Hi;
I get the same thing... Whenever my computer tries to upload any completed tasks, it receives the message: Message from server: Server error: can't attach shared memory.

I wonder if I did something? But I don't think so, as I have not changed anything in my prefs for either boinc manager or the web site.

Fix it Folks!

Larry

hey guys.... look here for a fix

Hitting 'update' about 10 times worked for me. PITA on 20 boxes, though.
I guess I could have done it thru boingmanager. Thanks moderators and others
for the suggestion.

/amgthis

Now getting a message from BOINC

Mon 01 Dec 2008 07:49:39 PM EST|rosetta@home|Message from server: Server error: can't attach shared memory

This has been happening the last couple of hours...

Ditto for me. And before that, I was seeing all of my Rosetta Mini 1.40 "abinitio_nohomgraf_..." tasks failing to report in for at least several hours this morning. They never did successfully transfer. I've got them queued up.

---

For those that are still having issues with your boinc client not getting the new scheduler url from our master url and if you would rather not detach and reattach the project which should fix the issue if updating doesn't. You can do the following:

1. stop the boinc client.
2. manually edit the client_state.xml, client_state_prev.xml, and master_boinc.bakerlab.org_rosetta.xml files in the BOINC client data directory by changing all occurrences of "https://boinc.bakerlab.org/rosetta_cgi/cgi" to "http://srv4.bakerlab.org/rosetta_cgi/cgi". This should be between <scheduler> and </scheduler> tags. You might not need to update master_boinc.bakerlab.org_rosetta.xml but you might as well.

If anyone has any suggestions or ideas for an easier fix, I'm all ears. The redirection does not appear to work so I removed it.

Does these suggestions lose all existing work? If so I really don't appreciate losing 4+ cpu hours per WU! Note this would be worst when you change the default to 6 hours.

it is only redirecting your communications/scheduler to the new correct server.
should have no impact on your tasks in process.
only do this if you are still having communications trouble.

THis worked for me.
Thanks!

---

We recently updated the web site with the latest BOINC code. Please post any issues with the web site in this thread. This update includes new features in the forum, spam filters, and a fix for merging hosts.

Any word on when the server memory problem will be fixed? I have two machines waiting.

Thanks

Lee Coin

Lee, your client will update itself in fairly short order. But you have to get it to attempt contact to the project by allowing more work. Then just leave it alone for a day and see what it does all by itself.

---

Rosetta Moderator: Mod.Sense

Suggested modification to the web site to avoid losing more participants:

Under Rosetta@home preferences, add the following feature, similar to what I've seen on other BOINC projects:

A number of selections for which workunits are allowed on this computer. For example:

1) test for workunits using a program that's passed RALPH but still needs some performance testing before being released to everyone. For now, I'd put workunits using the new features of minirosetta added in 1.39 and 1.40 in this class.

2) minirosetta for the other workunits using any version of minirosetta.

3) rosetta_beta for workunits using that program.

I'd set it up so all of them are enabled for users who haven't set any of them yet, except disable test for users who create accounts after this feature is added unless they then enable it. This should avoid any problems using computers where these settings haven't been made yet.

message board server is running really slow from about 00.10-00.15 UTC
it is taking 1.5-2 mins to load any one section of the page.
this slowness comes and goes at random
time now 00.16 UTC and it seems to run fine

I recently upgraded my version of BOINC from 5.10.45 to 6.2.28. Before that, I had selected 14 CPU hours as the preferred workunit length. Both of the Rosetta@home workunits I've downloaded since the BOINC upgrade came with an expected length of 29 hours. Are you out of 14 hour workunits? Or does Rosetta@home have problems with the combination of BOINC 6.2.28 and a selection of 14 hours as the preferred length?

1/1/2009 1:08:22 PM|rosetta@home|Sending scheduler request: To fetch work. Requesting 15375 seconds of work, reporting 0 completed tasks
1/1/2009 1:08:27 PM|rosetta@home|Scheduler request succeeded: got 1 new tasks
1/1/2009 1:08:29 PM|rosetta@home|Started download of mpzn_vvaa1t3kA09_05.200_v1_3.gz
1/1/2009 1:08:29 PM|rosetta@home|Started download of mpzn_vvaa1t3kA03_05.200_v1_3.gz
1/1/2009 1:08:52 PM|rosetta@home|Finished download of mpzn_vvaa1t3kA03_05.200_v1_3.gz
1/1/2009 1:08:52 PM|rosetta@home|Started download of mpzn_vv1t3kA.psipred_ss2.gz
1/1/2009 1:08:53 PM|rosetta@home|Finished download of mpzn_vv1t3kA.psipred_ss2.gz
1/1/2009 1:08:53 PM|rosetta@home|Started download of mpzn_vv1t3kA.pdb.gz
1/1/2009 1:08:55 PM|rosetta@home|Finished download of mpzn_vv1t3kA.pdb.gz
1/1/2009 1:09:45 PM|rosetta@home|Finished download of mpzn_vvaa1t3kA09_05.200_v1_3.gz


1/1/2009 2:47:43 PM|rosetta@home|Starting 1t3kA_BOINC_MPZN_vanilla_abrelax_5901_13195_1
1/1/2009 2:47:44 PM|rosetta@home|Starting task 1t3kA_BOINC_MPZN_vanilla_abrelax_5901_13195_1 using minirosetta version 147
1/1/2009 2:47:46 PM|rosetta@home|Sending scheduler request: To fetch work. Requesting 3208 seconds of work, reporting 0 completed tasks
1/1/2009 2:47:51 PM|rosetta@home|Scheduler request succeeded: got 1 new tasks
1/1/2009 2:47:53 PM|rosetta@home|Started download of boinc_mfr_aawd20_03_05.200_v1_3.gz
1/1/2009 2:47:53 PM|rosetta@home|Started download of boinc_mfr_aawd20_09_05.200_v1_3.gz
1/1/2009 2:48:01 PM|rosetta@home|Finished download of boinc_mfr_aawd20_03_05.200_v1_3.gz
1/1/2009 2:48:01 PM|rosetta@home|Started download of wd20_.fasta
1/1/2009 2:48:02 PM|rosetta@home|Finished download of wd20_.fasta
1/1/2009 2:48:02 PM|rosetta@home|Started download of boinc_description_file.txt
1/1/2009 2:48:03 PM|rosetta@home|Finished download of boinc_description_file.txt
1/1/2009 2:48:03 PM|rosetta@home|Started download of wd20.pdb
1/1/2009 2:48:06 PM|rosetta@home|Finished download of wd20.pdb
1/1/2009 2:48:06 PM|rosetta@home|Started download of wd202.pdb
1/1/2009 2:48:09 PM|rosetta@home|Finished download of wd202.pdb
1/1/2009 2:48:16 PM|rosetta@home|Finished download of boinc_mfr_aawd20_09_05.200_v1_3.gz

Mod.Sense suggested a quick fix for this by using the Redirect directive on our webserver. The main question now is does the client handle redirects so please let me know if this fixes the issue or not.

Thanks Mod.Sense!

Note that I've seen messages of other BOINC projects indicating that if you make certain updates to the selection of servers, just one update is not enough to make clients download the file specifying which servers to use; instead, it takes about 10 updates in a row that try to download at least one workunit, but don't succeed in downloading any at all.