Hi. I added, a few days ago a new machine to crunch Rosetta ( as well as a few other projects) on Rosetta it seems that everytime this machine connects it generates a new host ID. On one of the BOINC boards I frequent it was suggested that this had something to do with the version of software running at the host site. Is there any way this could be checked?

Thanks

Steve

---

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

Thanks! I don't think we can merge right now in Rosetta though.

---

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

Thanks! I don't think we can merge right now in Rosetta though.

There is a known bug in the BOINC BAM system that caused this problem. IF you stop using the BAM the problem will stop by itself

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

Thanks! I don't think we can merge right now in Rosetta though.

True I'm afraid! My machine contacts Rosetta once an hour and creates a new host each time. It's doing the same on LHC, but at least they are allowing us to merge hosts.

Anyone know when the merge facility will be turned back on?

---

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

Thanks! I don't think we can merge right now in Rosetta though.

There is a known bug in the BOINC BAM system that caused this problem. IF you stop using the BAM the problem will stop by itself

It is nothing to do with BAM it is a server side problem that creates this error a patch is there and ready to be used but ROSETTA won't do anything until CASP7 finishes.

The error arrived on various projects before BAM came on the scene

---

Join us in Chat (see the forum) Click the Sig


Join UBT

Thanks! I don't think we can merge right now in Rosetta though.

True I'm afraid! My machine contacts Rosetta once an hour and creates a new host each time. It's doing the same on LHC, but at least they are allowing us to merge hosts.

Anyone know when the merge facility will be turned back on?

The best information I have is that it is highly unlikely that a server software upgrade will be attempted during CASP. If for some reason there were a break in the action then maybe. But the project does not want to risk the projects position to fix a problem that while very inconvenient, is not actually critical to the generation of results. Please bear with them on this. If they lost a day of down time they would not be able to generate sufficient models to complete the CASP work.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

Thanks! I don't think we can merge right now in Rosetta though.

There is a known bug in the BOINC BAM system that caused this problem. IF you stop using the BAM the problem will stop by itself

Stopping BAM doesn't help. I tried that earlier in the week and it did no good. I think UBT - Halifax--lad has the most correct info on this right now from what I've read in other forums.

Thanks Moderator9 and UBT - Halifax--lad for all your help. I guess I'll just keep crunching and let things sort themselves out. I'll keep my fingers crossed for a quick break in the action that will allow the patch to be applied. I certainly understand any reluctance to modify things while CASP is in progress.

---

The bug is listed. A new client version is coming soon. I solved this part by merging the last host entry with others. No new host was created.

See also here: http://boinc.gorlaeus.net/forum_thread.php?id=99

Thanks! I don't think we can merge right now in Rosetta though.

There is a known bug in the BOINC BAM system that caused this problem. IF you stop using the BAM the problem will stop by itself

It is nothing to do with BAM it is a server side problem that creates this error a patch is there and ready to be used but ROSETTA won't do anything until CASP7 finishes.

The error arrived on various projects before BAM came on the scene

What are the conditions that are required for this to happen every time a Boinc project is contacted - outside of BAM usage?

---

...
What are the conditions that are required for this to happen every time a Boinc project is contacted - outside of BAM usage?

For some reason the BOINC server thinks that each connection represents a new attachment to the project. Each attach function creates a new computer ID. The cause is some sort of signal mismatch between the BOINC client and the server. The problem is not universal, and as far as anyone knows, only BAM users are seeing it at all.

I have sent a report on this thread to the system admin at Rosetta. I am sure he will check on this from his end again (this is not the first report). But There will have to be a very compelling reason to shut down the servers in the middle of CASP with a number of protein reporting deadlines scheduled for Tuesday.

At least two users solved this issue by not using BAM, resetting the project, and restarting BOINC.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

As mentioned, if you read the fine print at the BAM website, it does not list R@H as a BAM compatible project....Cheers, Rog.

---

BAM may escalate the problem due to connecting every hour but it is not the cause of the problem the cause of the problem is an error within BOINC servers & the client, the client has had a fix put into it but for it to work the server also needs its fix. As soon as the fix is in multiple hosts will stop been created.

It has been a known problem with BOINC for many months now, it 1st appeared on SETI where they corrected the error straight away with there patch, most projects are now upgrading there servers with the patch (well the ones its been reported on).

Due to CASP though we will have to sit and wait here for it to end

---

Join us in Chat (see the forum) Click the Sig


Join UBT

Just a thought...

I have two machines, one with no location set and one with the location set to "work" in order to have one working all the time and the other only when the user is not active. The machine with no location set is not creating new host IDs but the one with the "work" location is.

I originally set the "work" location up on BAM, but it seems not to create a new "work" location on any projects (not just Rosetta) where no "work" location had been set up beforehand. I have now manually created a "work" location on each of the offending projects (Rosetta, LHC and Predictor). However, this doesn't seem to have helped, as "new" host IDs are still being created, but with a location of "home".

Would it help to delete the "work" location and all references to it, and set the location for the second machine to "home" or to no location? I guess that if I wanted to set no location I'd have to either detach that machine from BAM, or else maybe set an unrecognised location for that machine on BAM to stop it resetting back to "home". The first machine was originally set up pre-BAM with no location, and I created a "home" location on BAM, with the result that all projects (including Rosetta) now have a location of "home" with no problems. Hmmmm.....

---

Well, I'm NOT using BAM at the moment and I also have 14 entries of the same host. That's ok unless I want to start crunching again for Rosetta. It would of course be very annoying if every scheduler contact adds one entry more to the list... ;)

Edit: now after detaching and re-attaching I have host entry no #15 *LOL*
In the messages tab of the BM I have the usual line:

27.05.2006 15:21:35|rosetta@home|URL: https://boinc.bakerlab.org/rosetta/; Computer ID: not assigned yet; location: home; project prefs: default

---

Grrrrrrrrreetings from the Lazy Cat...

BAM may escalate the problem due to connecting every hour but it is not the cause of the problem the cause of the problem is an error within BOINC servers & the client, the client has had a fix put into it but for it to work the server also needs its fix. As soon as the fix is in multiple hosts will stop been created.

Well...

Ok, if Rosetta@home hadn't been running outdated Scheduling-server, you wouldn't generate a new hostid per connection, this part is true...

... But, if user haven't changed Resource-share in any Account Manager, you won't have the problem of a new hostid being generated for each connection, even if a project runs outdated Scheduling-server...

Due to CASP though we will have to sit and wait here for it to end

And while waiting, a very easy work-around, just edit the resource-share on the individual projects web-sites, instead of in the Account Managers...


BTW, if all your computers/projects does have hostid, you can safely change resource-shares in Account Manager. Also, atleast Einstein@home works even if you don't have hostid from before, not sure if any other project works correctly, and for obvious reasons can't check at the moment...

Not prepared to do anything will wait for Rosetta to update there server

---

Join us in Chat (see the forum) Click the Sig


Join UBT

I already have 60 hosts (from 8 or 10 yesterday) and climbing. :| But this problem only aperas on the newly installed machines, the old ones are working just fine. At least, after the software update, we might be able to merge the hosts.

---

hey, how to get those banners.. i am also running rosetta for most of the time, i also want to add one such banner.

---

hey, how to get those banners.. i am also running rosetta for most of the time, i also want to add one such banner.

Here's a link for you to one of the stats sites that can give you a signiture line.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

hey, how to get those banners.. i am also running rosetta for most of the time, i also want to add one such banner.

Here's a link for you to one of the stats sites that can give you a signiture line.

Or you can go to Mundayweb.com need to create an account there

or mine came from UBT Sig

---

Join us in Chat (see the forum) Click the Sig


Join UBT