Distro questionnaire

Please report the Linux distro(s) you are running here and the fact whether all of your wu's are erroring out on it.

Thanks
Andre
____________
D@H the greatest project in the world... a while from now!

Please report the Linux distro(s) you are running here and the fact whether all of your wu's are erroring out on it.

Thanks
Andre

Please report the Linux distro(s) you are running here and the fact whether all of your wu's are erroring out on it.

Thanks
Andre

Ubuntu 5.10, all giving the exit code 1.

PCLinuxos 0.93a "The full Monty"{Big Daddy}
The PCLinuxOS 0.93a Full Edition was quietly released yesterday and is the equivalent of PCLinuxOS releases of the past several years. They have recently been releasing scaled down versions to accomodate other tastes and desires, but many folks were looking and waiting for the fully loaded edition. Weighing in at 685 MB, Texstar referred it as the "Full Monty." Its code name has been "Big Daddy." Whatever the name, I think you'll call it home.

Before my daily limit was reached the new app errored out
Mon Sep 18 17:24:12 2006|Docking@Home|Unrecoverable error for result 1tng_mod0001_1378_275472_2 (process exited with code 1 (0x1))

FWIW, my Linux systems (SUSE ES 9.0 and SUSE 10.0) started runnning successfully (e.g., this one ) a while ago.

HTH

____________

Fedora Core 3 with all updates. All WUs have errored out so far. It's just now downloading version 5.2 of the charmm app. Today was my first day so the errored out units (which seem to have reported success) were running whatever version of the docking app you were using this morning. Anyway, thay all got the following result after just under 300 seconds.

<core_client_version>5.4.9</core_client_version>
<stderr_txt>
Starting charmm run...
ERROR - Charmm exited with code 1.
Calling BOINC finish.

</stderr_txt>

The machine is a Socket A sempron 2500+ with 1 GB ram and nothing but BOINC running on it. It has several GB of disk space available. hostid=216

-- David

>> I am using Fedora Core 3, all updates installed on an AMD Opteron 848 dual cpu.
All work units (between 200 and 300 that have been processed), have errored out with Error code 1.
All Units have taken about 3 1/2 minutes.
____________

I am using Debian Etch with all the updates as of yesterday (9-17-06). I have two machine both of which are running this same distro. I just joined today and both boxes downloaded 5.01. One machine is running and returning results without any problems. The other is giving me errors.

Date Message
9/18/2006 2:32:04 PM Starting task 1tng_mod0001_1161_91144_2 using charmm version 501
9/18/2006 2:32:06 PM Unrecoverable error for result 1tng_mod0001_1161_91144_2 (process got signal 4)
9/18/2006 2:32:06 PM Deferring scheduler requests for 1 minutes and 0 seconds
9/18/2006 2:32:06 PM Computation for task 1tng_mod0001_1161_91144_2 finished
9/18/2006 2:33:10 PM Sending scheduler request to http://docking.utep.edu/docking_cgi/cgi


I reached my daily quota of 25 on this machine. I've reset it, but will have to wait until tommorrow to see if I get the 5.02 app.

All projects on this machine are running fine, with the exception of this project and HashClash.

Jim

All WUs crashing here too, running Fedora Core 5 on 4 older Intel and 1 AMD64 2.4 GHz and all are the same, all crash. A few units report Success on the stats page but even those turn out to be crashed units when I examine the crunch report.

Jim,

signal 4 is a termination of the process because of an illegal instruction. This can happen with any process on linux/unix incl. our app. I think you're the first that has seen one of those. Of course it might be that somebody else has seen this and didn't report it. Please try again tomorrow when it downloads 5.2.

Thanks
Andre

PS The daily quota is set to 50 for D@H

I am using Debian Etch with all the updates as of yesterday (9-17-06). I have two machine both of which are running this same distro. I just joined today and both boxes downloaded 5.01. One machine is running and returning results without any problems. The other is giving me errors.

Date Message
9/18/2006 2:32:04 PM Starting task 1tng_mod0001_1161_91144_2 using charmm version 501
9/18/2006 2:32:06 PM Unrecoverable error for result 1tng_mod0001_1161_91144_2 (process got signal 4)
9/18/2006 2:32:06 PM Deferring scheduler requests for 1 minutes and 0 seconds
9/18/2006 2:32:06 PM Computation for task 1tng_mod0001_1161_91144_2 finished
9/18/2006 2:33:10 PM Sending scheduler request to http://docking.utep.edu/docking_cgi/cgi


I reached my daily quota of 25 on this machine. I've reset it, but will have to wait until tommorrow to see if I get the 5.02 app.

All projects on this machine are running fine, with the exception of this project and HashClash.

Jim

Still working on this problem with the exit 1 crashes. This is the hardest problem we've had up to now, but hopefully we'll get closer to a solution tomorrow.

If linux users want to suspend until we got this fixed, please do so.

Thanks
Andre

All WUs crashing here too, running Fedora Core 5 on 4 older Intel and 1 AMD64 2.4 GHz and all are the same, all crash. A few units report Success on the stats page but even those turn out to be crashed units when I examine the crunch report.

It seems that SuSE users seem to be lucky with this one.
Please keep on crucnhing if you are running SuSE, but feel free to suspend for now if you are errorring out on all the workunits you receive.

Andre

FWIW, my Linux systems (SUSE ES 9.0 and SUSE 10.0) started runnning successfully (e.g., this one ) a while ago.

HTH

All WUs are failing on RHEL3 on a Celeron 2.4 Ghz, 2 GB ECC Ram, several GB of disk space. This is for both Charmm 5.01 ( which returned success but really errored out) and 5.02 which returns client error/compute error.

<core_client_version>5.2.13</core_client_version>
<message>process exited with code 1 (0x1)
</message>
<stderr_txt>
Calling BOINC init.
Starting charmm run...
ERROR - Charmm exited with code 1.
Calling BOINC finish.

</stderr_txt>

CPU time ranges from 387 seconds to 405 seconds.

hostid = 223
Linux kernel 2.4.21-47.EL

5.02 is no different, still getting the same Error code 1 failure. Still 3 1/2 minutes per WU and none successful. Fedora Core 3. Will suspend as you, Andre, have instructed.
____________

It seems that SuSE users seem to be lucky with this one.
Please keep on crucnhing if you are running SuSE, but feel free to suspend for now if you are errorring out on all the workunits you receive.

Andre

We have finally found the cause of the problem that some users were experiencing on their Linux systems. It has to do with the stacksize setting on your machine which is for some distros (SuSE 9.3 and 10 for example) set to unlimited and for others (FCx, Ubuntu, etc) set to a limited value like 10240. Your setting can be seen by typing 'ulimit -s' in a terminal. To make the Charmm 'exit 1' errors go away, please set the stacksize to unlimited using the command 'ulimit -s unlimited'. This is not saying that Charmm will use all of your memory (it won't), but it gives us a little bit more space to do our simulations correctly and without errors. Please let us know if this does not work for you. If it does work, please add this command to your shell initialization file (.bashrc, .tcshrc, .kshrc, etc) in your home directory. Of course don't forget to resume the D@H project on your boincmgr in case you suspended it before.

Thanks,
Andre

It seems that SuSE users seem to be lucky with this one.
Please keep on crucnhing if you are running SuSE, but feel free to suspend for now if you are errorring out on all the workunits you receive.

Andre

I use SUSE 10.1 and all of my 125 results have the error code 1 with the 5.01 app. Havn´t tried 5.02, but you said there isn´t a fix for the error 1 in it... :(

It seems that SuSE users seem to be lucky with this one.
Please keep on crucnhing if you are running SuSE, but feel free to suspend for now if you are errorring out on all the workunits you receive.

Andre

Good news !

My Mandrake distro said "8192" for ulimit -s.
I changed to "unlimited".

Wait and see ...
____________

Thanks Andre,
I did the "ulimit -s" then "ulimited -s" with the Linux console ran 2 w/u and received code 1 error
Tue Sep 19 16:15:56 2006|Docking@Home|Unrecoverable error for result 1tng_mod0001_3688_167242_2 (process exited with code 1 (0x1))
The last application worked for me albeit with the error code.What other info can help you ?Is there another way to fix this Andre?
Thanking you in advance


Doug Worrall

typed in "su" and did the ulimit -s within root.Then tryed in user and got this number 8192, ran 2 w/u and errored out.Looks as though I am out of options, unless
there are the older w/u that worked.I forgot 5.01?.?
Andre, if possible, could the 200% time share post Please be deleted?

Thanking you in advance

Doug Worrall

I'm running Fedora Core 5, ulimit -s gave me 8192, changed it to unlimited and I'm still getting exit code 1 on all my WU's.

Di 19 Sep 2006 23:45:19 CEST|Docking@Home|Unrecoverable error for result 1tng_mod0001_3744_201305_2 (process exited with code 1 (0x1))

Please check this thread It's the same problem. Error results get successfully validated and successful results do not. As soon as all 5.1 results are out of the system, this problem should be gone.

Andre

It seems that SuSE users seem to be lucky with this one.
Please keep on crucnhing if you are running SuSE, but feel free to suspend for now if you are errorring out on all the workunits you receive.

Andre

I've running Suse 9.3. My WU's are successful since version 5.02, but the running time of some WU's are very different from others who run the same WU and then the result is invalid.

1) Open a terminal
2) Type 'ulimit -s unlimited' in that terminal (without the quotes)
3) Change directory to your BOINC installation and type './run_manager' in that same terminal.

That should work.

What's that 200% time share post you mentioned?

AK

Thanks Andre,
I did the "ulimit -s" then "ulimited -s" with the Linux console ran 2 w/u and received code 1 error
Tue Sep 19 16:15:56 2006|Docking@Home|Unrecoverable error for result 1tng_mod0001_3688_167242_2 (process exited with code 1 (0x1))
The last application worked for me albeit with the error code.What other info can help you ?Is there another way to fix this Andre?
Thanking you in advance


Doug Worrall

See my previous post

Andre

I'm running Fedora Core 5, ulimit -s gave me 8192, changed it to unlimited and I'm still getting exit code 1 on all my WU's.
Di 19 Sep 2006 23:45:19 CEST|Docking@Home|Unrecoverable error for result 1tng_mod0001_3744_201305_2 (process exited with code 1 (0x1))

I'm trying to delete this post but it looks like the best I can do is edit it.

1) Open a terminal
2) Type 'ulimit -s unlimited' in that terminal (without the quotes)
3) Change directory to your BOINC installation and type './run_manager' in that same terminal.

I am very new to Linux.
In a terminal screen I have typed 'ulimit -s' got 10240. I then typed 'ulimit -s unlimited' and hit enter on the Terminal screen. This did seem to do much but a further 'ulimit -s' gave back 'unlimited' as the stack size. So I downloaded some WU's but after 3 1/2 minutes they all errored out.
If the 'unlimited' thing works (which I don't see any evidence of) how do I put that line anywhere so that it stays and gets initialized with the running of Boinc?
I don't understand line 3, probably because I have no listing of commands and Linux terminal screen operation.
Can someone give me more detailed instructions please?
If I am unable to get this computer to process anything I may as leave as I am not being of much help at the moment. It does noy matter if I use 5.01 or 5.02 as all units go 3 1/2 minutes and error out.
Sorry for being a pain, but I would like to help.
____________

Hi Conan,

Find a file called run_manager in your BOINC installation directory and add the line 'ulimit -s unlimited' (without the quotes) as the first line in that file using your favorite editor. Now stop your running boinc manager, open a terminal, type 'cd <boinc installation directory> followed by ./run_manager. You might also be able to use your GUI file manager to go to that directory and double-click on the run_manager file.

Let me know if that works.
Andre

1) Open a terminal
2) Type 'ulimit -s unlimited' in that terminal (without the quotes)
3) Change directory to your BOINC installation and type './run_manager' in that same terminal.

I am very new to Linux.
In a terminal screen I have typed 'ulimit -s' got 10240. I then typed 'ulimit -s unlimited' and hit enter on the Terminal screen. This did seem to do much but a further 'ulimit -s' gave back 'unlimited' as the stack size. So I downloaded some WU's but after 3 1/2 minutes they all errored out.
If the 'unlimited' thing works (which I don't see any evidence of) how do I put that line anywhere so that it stays and gets initialized with the running of Boinc?
I don't understand line 3, probably because I have no listing of commands and Linux terminal screen operation.
Can someone give me more detailed instructions please?
If I am unable to get this computer to process anything I may as leave as I am not being of much help at the moment. It does noy matter if I use 5.01 or 5.02 as all units go 3 1/2 minutes and error out.
Sorry for being a pain, but I would like to help.

The "ulimit -s unlimited" fixed the problem on Fedora Core 3. The workunit took just over 2 hours, Charmm exited with code 0, returned success, and validated.

App version 5.02
Granted credit 9.92

-- David Ball

>> Thanks Andre, that has worked and such a simple looking solution. I have now processed 5 successful Work Units that have all taken about 5213 seconds or 1.45 hours (1 hour 26 minutes).
They show "Success with exit code 0".
Solving this problem should convince the boss to let you have some time to build your kit plane and "de-stress".
Thanks again for the quick response.
____________

>>Solving this problem should convince the boss to let you have some time to build your kit plane and "de-stress".

OK. I type and change step by step and nothing!
The same as before ... crash.
I stopped the client.
(Kubuntu 606 , AMD Athlon 2800+ )

Just to be sure... you did:

1) stop any running boinc clients/managers
2) open a terminal and type 'echo $SHELL' to find out which shell you are running (for this and future commands: do not type the 'quotes')
3) if it says 'bash' type 'ulimit -s unlimited'
4) change to the directory where you installed boinc (e.g. 'cd ~/BOINC' if its in your home)
5) type './boincmgr' to start the boinc manager
6) runs some jobs and see whether they look allright this time

Andre

OK. I type and change step by step and nothing!
The same as before ... crash.
I stopped the client.
(Kubuntu 606 , AMD Athlon 2800+ )

matze@Kubuntu-der-Erste:~$ echo $SHELL
/bin/bash
matze@Kubuntu-der-Erste:~$ ulimit -s unlimited
matze@Kubuntu-der-Erste:~$ ulimit -s
unlimited
matze@Kubuntu-der-Erste:~$
...
matze@Kubuntu-der-Erste:~/Cruncher/BOINC$ ./run_manager
send: -1
send: Bad file descriptor
connect: Connection refused
connect: Operation now in progress

(In the terminal.)

And now 5min + + + It`s running. 6min + + +

Thanks, thanks ... :-)

Kubuntu Dapper 6.06 working greatly with ulimit fix, both on intel (Core duo T2500) and amd (Athlon XP) processors
(hd writes rating still a bit high)

Kubuntu Dapper 6.06 working greatly with ulimit fix, both on intel (Core duo T2500) and amd (Athlon XP) processors
(hd writes rating still a bit high)

VineLinux (on VMWare) works well with ulimited -s unlimited:)
I started testing of the distro since yesterday, so I report so later than other members...

Edit: Now I'm testing "debian for VMWare" on the virtual machine. I've already modified run_manager and run_client. I'll report issues if exist:)
____________

I'm a volunteer participant; my views are not necessarily those of Docking@Home or its participating institutions.

Ubuntu 6.10
Linux 2.6.17-10
Boinc client and manager 5.4.11-1

This distro has no run_manager or run_client
Added ulimit -s unlimited to boinc-client a script file located in /etc/init.d

This seems to have fixed the problem. I'm running Linux on a slow box (Celeron 1200 MHz, 384 mem). It took me a little over 8.5 hours to complete the first work unit. Working on another one now.

Damn Small Linux runs fine with no need for the ulimit fix, even though the stack is a default 8192, although all my boxes running this distro are PII/PIII variations and failing validation.
____________

Please make sure that there has been a issue with PIII unsolved yet.
____________

I'm a volunteer participant; my views are not necessarily those of Docking@Home or its participating institutions.

This is mainly due to the biggest problem we are working on currently: our app returns different results on different platforms, even among different Intel cpu's! Fixing this is our main focus at the moment; most other issues have to take a back seat for the moment.

Thanks
Andre

Damn Small Linux runs fine with no need for the ulimit fix, even though the stack is a default 8192, although all my boxes running this distro are PII/PIII variations and failing validation.

red hat fedora core 2.6
ulimit -s 10240
machine hangs, wu failed on restart avg 7 minutes

set to unlimited
machine hangs
rebooted it continues
the last wu hung me 6 times in 2:53 @approx 72 % then was erred on restart

I suspect hardware, as only xtrem chess and pirates, dont hang. :(

am try again (I cant find ANY other app to crash that pc lol not memtest or any 'burn in app', and just hanging like that i get no crash reports lol )

was going to suse this pc, athlon64, last week but the spindle of cdr's i got fail to record properly. kicking around distros' ez, ark, debian, gentoo, and red hat all failed to install for 1 reason or another (8Mb pci gc //sata raid // gigalan) too old and too new

i ran ez on my intel for a few days (You had no wu 4 me) it only found 1 core benchmarked poor (got good credit at rosetta tho)
the opteron i fixxed and played with for a day was 2.6.16-xenU ran tampaku cuz U weren't here.

i tried the distros on my 'hammer' AMD but the net card is 'wondoze' dependent not supported due to liscencing restrictions and i got doing something else.

think i'll play with the 'use xx% memory or processer setting and see if the old 2100 can succeed 1 wu at least lol

sorry for being wordy, just letting You know I'm here and trying all my oddball junky old machines, I'll let You know.
____________

i see 1 wu succeeded (on the 14th)(went out for diner, the pc was hung when i got home but it musta run a whole wu) am getting the 'other platforms message' will drop by when i wake up
____________