SERVER PROBLEMS.

Message boards : Number crunching : SERVER PROBLEMS.

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 . . . 12 · Next

AuthorMessage
P . P . L .

Send message
Joined: 20 Aug 06
Posts: 581
Credit: 4,865,274
RAC: 0
Message 57440 - Posted: 2 Dec 2008, 2:41:15 UTC

Hi.

If you are using windows, you could try this i found it on seti forums

might be worth a go. (Don't know if it will i use Linux.)

"Can't resolve hostname" is usually caused by an invalid cached dns entry. Try opening a command prompt (if you're using Vista, make sure you run the command prompt as Administrator) and type in:

IPCONFIG /flushdns

pete.


ID: 57440 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Otto

Send message
Joined: 6 Apr 07
Posts: 27
Credit: 3,567,665
RAC: 0
Message 57442 - Posted: 2 Dec 2008, 2:53:37 UTC

Oh, oh, please don't say that I will lose several WUs of 6 hour crunches. I wouldn't be too pleased about it (if I have to detach the project then re-attach in order to make it work). Refreshing the project did nothing.
ID: 57442 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 57444 - Posted: 2 Dec 2008, 3:32:02 UTC

The "Message from server: Server error: can't attach shared memory" problem seems to be with the Rosetta servers. My machines were able to download work, etc for a few hours today, but now they get that error message.

I remember this problem happening in the past (over a year ago), but I can't remember what it turned out to be. Was it that some piece of shared memory in the server programs wasn't big enough? Can anyone else remember?

ID: 57444 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile David E K
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 1 Jul 05
Posts: 1018
Credit: 4,334,829
RAC: 0
Message 57445 - Posted: 2 Dec 2008, 3:37:43 UTC

You should be able to just update the project by clicking the update button on the boinc manager when the R@h project is selected. This should cause the client to get the new scheduler url from the master url.

The scheduler changed from https://boinc.bakerlab.org/rosetta_cgi/cgi to http://srv4.bakerlab.org/rosetta_cgi/cgi

The client should get this information from https://boinc.bakerlab.org/rosetta. The only reason I can see it not working is if this url is cached somehow and the client is getting an older (earlier today) version.

let me know if you continue to have problems.
ID: 57445 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile David E K
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 1 Jul 05
Posts: 1018
Credit: 4,334,829
RAC: 0
Message 57447 - Posted: 2 Dec 2008, 4:09:07 UTC

Mod.Sense made an excellent suggestion for a quick fix to this which involved making our server redirect requests to our old scheduler url to the new one. If the client is smart enough to use the redirect, this should fix things right away. Let me know if the problem persists.
ID: 57447 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 57448 - Posted: 2 Dec 2008, 4:22:44 UTC

When I press "update" I now get the message "Message from server: Incomplete request received."
ID: 57448 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile David E K
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 1 Jul 05
Posts: 1018
Credit: 4,334,829
RAC: 0
Message 57450 - Posted: 2 Dec 2008, 4:47:15 UTC - in response to Message 57448.  

When I press "update" I now get the message "Message from server: Incomplete request received."


Looks like the client isn't following the redirect. That's unfortunate. I'll look into it further.
ID: 57450 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile XJR-Maniac
Avatar

Send message
Joined: 24 Jan 06
Posts: 12
Credit: 271,462
RAC: 0
Message 57455 - Posted: 2 Dec 2008, 7:55:46 UTC

Still getting this message:


Scheduler RPC succeeded
Message from server: Server error: can't attach shared memory
Deferring communication for 1 hr 0 min 0 sec
Reason: project is down


Did a flushdns, stop/start of boinc service, no change.

ID: 57455 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
TomaszPawel

Send message
Joined: 28 Apr 07
Posts: 54
Credit: 2,791,145
RAC: 0
Message 57456 - Posted: 2 Dec 2008, 8:52:14 UTC - in response to Message 57455.  

2008-12-02 09:45:55|rosetta@home|Message from server: Server error: can't attach shared memory

What this mean?
ID: 57456 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Matthew Lei
Avatar

Send message
Joined: 5 Jun 06
Posts: 4
Credit: 258,058
RAC: 0
Message 57459 - Posted: 2 Dec 2008, 9:37:39 UTC

looks like many of us are having this shared memeory issue. Lets hope that admins are going to fix it soon.
ID: 57459 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 57461 - Posted: 2 Dec 2008, 10:05:31 UTC
Last modified: 2 Dec 2008, 10:23:11 UTC

I'm leaving my machines alone, and after about 6 hours and many automatic retries, they eventually say "Fetching master file" and start working. (I'm using BOINC 5.2.13 on these machines.)
ID: 57461 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Evan

Send message
Joined: 23 Dec 05
Posts: 268
Credit: 402,585
RAC: 0
Message 57462 - Posted: 2 Dec 2008, 10:08:07 UTC

You can fix it yourself. Their instructions worked for me.
The problem is because they have moved to another server to ease the load.
Exit boinc and then
Open notepad and set it for all files
and then open the boinc file - in my case found in documents and settings/all users/application data/boinc.

There are three files to edit
client_state.xml,
client_state_prev.xml,
master_boinc.bakerlab.org_rosetta.xml

Open them in turn and enter "schedule" in the edit/find tab of notebook.
This will take you to the correct line.

<scheduler_url>http://bakerlab.org/rosetta_cgi/cgi</scheduler_url>
enter srv4. before the bakerlab.....
it will look like

<scheduler_url>http://srv4.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

repeat for the remaining two.

Restart boinc and update rosetta
ID: 57462 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Erwin Schlonz
Avatar

Send message
Joined: 20 May 07
Posts: 5
Credit: 203,397
RAC: 0
Message 57467 - Posted: 2 Dec 2008, 11:58:10 UTC - in response to Message 57462.  
Last modified: 2 Dec 2008, 12:56:50 UTC

You can fix it yourself.
There are three files to edit
client_state.xml,
client_state_prev.xml,
master_boinc.bakerlab.org_rosetta.xml

<scheduler_url>http://bakerlab.org/rosetta_cgi/cgi</scheduler_url>
<scheduler_url>http://srv4.bakerlab.org/rosetta_cgi/cgi</scheduler_url>
repeat for the remaining two.
Restart boinc and update rosetta


This worked for me. But the uploaded results now have the status pending.
Has this something do to with this fix?

[EDIT]
Ok, forget it. The results are now validated.
ID: 57467 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
The Grinch

Send message
Joined: 29 Mar 07
Posts: 3
Credit: 3,622,517
RAC: 0
Message 57468 - Posted: 2 Dec 2008, 12:54:21 UTC
Last modified: 2 Dec 2008, 12:54:45 UTC

02.12.2008 13:39:32|rosetta@home|Message from server: Server error: can't attach shared memory
02.12.2008 13:39:35||Suspending computation - user request

Is there any Server-Problems?
ID: 57468 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1981
Credit: 38,436,301
RAC: 14,059
Message 57470 - Posted: 2 Dec 2008, 14:00:12 UTC - in response to Message 57462.  
Last modified: 2 Dec 2008, 14:01:49 UTC

You can fix it yourself. Their instructions worked for me.
The problem is because they have moved to another server to ease the load.

Exit boinc and then
Open notepad and set it for all files
and then open the boinc file - in my case found in documents and settings/all users/application data/boinc.

There are three files to edit
client_state.xml,
client_state_prev.xml,
master_boinc.bakerlab.org_rosetta.xml

Open them in turn and enter "schedule" in the edit/find tab of notebook.
This will take you to the correct line.

<scheduler_url>http://bakerlab.org/rosetta_cgi/cgi</scheduler_url>
enter srv4. before the bakerlab.....
it will look like

<scheduler_url>http://srv4.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

repeat for the remaining two.

Restart boinc and update rosetta

Thanks for this. I used the IPCONFIG /flushdns instruction as well and it didn't work.

I was about to post until I realised the URL in those files was actually:

<scheduler_url>http://boinc.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

...so when I put srv4 in front of it it came out srv4.boinc.bakerlab.org instead of srv4.bakerlab.org

Stopping the BOINC service, a quick edit, restarting the BOINC service, resulted in downloading a new master file schedule list and all seems to be well.

Now to reset the project, allow new tasks and hope I'm back on stream again. Phew!

Thanks for everyone's help...

EDIT: And 9 new WUs come straight down. Let's hope this resolves a few issues now after this cleanup
ID: 57470 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Evan

Send message
Joined: 23 Dec 05
Posts: 268
Credit: 402,585
RAC: 0
Message 57473 - Posted: 2 Dec 2008, 14:43:52 UTC

Quite right, I didn't see that error. As you say it should read:

<scheduler_url>http://srv4.boinc.bakerlab.org/rosetta_cgi/cgi</scheduler_url>
ID: 57473 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
BarryAZ

Send message
Joined: 27 Dec 05
Posts: 153
Credit: 30,843,285
RAC: 47
Message 57495 - Posted: 2 Dec 2008, 17:32:50 UTC - in response to Message 57473.  

A simpler brute force approach (which will lose completed unreported work, and work in process) would be to detach and rejoin.


Quite right, I didn't see that error. As you say it should read:

<scheduler_url>http://srv4.boinc.bakerlab.org/rosetta_cgi/cgi</scheduler_url>


ID: 57495 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Evan

Send message
Joined: 23 Dec 05
Posts: 268
Credit: 402,585
RAC: 0
Message 57497 - Posted: 2 Dec 2008, 17:43:56 UTC
Last modified: 2 Dec 2008, 17:44:38 UTC

There is always a simpler way. The latest is to speed up the boinc process of updating the masterfile as found by AMD is logical

message 4539
ID: 57497 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 1981
Credit: 38,436,301
RAC: 14,059
Message 57502 - Posted: 2 Dec 2008, 18:32:42 UTC - in response to Message 57473.  

Quite right, I didn't see that error. As you say it should read:

<scheduler_url>http://srv4.boinc.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

No, you were right with the correction to make (coz it didn't work for me first time), but for me it's from:

<scheduler_url>https://boinc.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

should be changed to:

<scheduler_url>http://srv4.bakerlab.org/rosetta_cgi/cgi</scheduler_url>

So, replace "boinc" with "srv4" rather than just inserting svr4 in front of the full previous URL, if you get me.

That's what did it for me anyway
ID: 57502 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Heidi1
Avatar

Send message
Joined: 11 Aug 07
Posts: 49
Credit: 1,786,248
RAC: 0
Message 57505 - Posted: 2 Dec 2008, 20:05:25 UTC - in response to Message 57462.  

You can fix it yourself. Their instructions worked for me.
The problem is because they have moved to another server to ease the load.
Exit boinc and then
Open notepad and set it for all files
and then open the boinc file - in my case found in documents and settings/all users/application data/boinc.

There are three files to edit
client_state.xml,
client_state_prev.xml,
master_boinc.bakerlab.org_rosetta.xml

Open them in turn and enter "schedule" in the edit/find tab of notebook.
This will take you to the correct line.

<scheduler_url>http://bakerlab.org/rosetta_cgi/cgi</scheduler_url>
enter srv4. before the bakerlab.....
it will look like

<scheduler_url>http://srv4.bakerlab.org/rosetta_cgi/cgi</scheduler_url>


I followed the instructions, with both versions (the one quoted above and the version with srv4.boinc.bakerlab...), and neither worked. I now have an error message saying that there is no start tag in the server reply and that it can't parse the file. I don't know what that means, and what I can do to fix it, other than going to my backup of the files, copying them back into my computer, and hitting the Update button 11 times. Any thoughts?

ID: 57505 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 · 2 · 3 · 4 · 5 . . . 12 · Next

Message boards : Number crunching : SERVER PROBLEMS.



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