Global Correlations S5 - Signal 11

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0
Topic 194908

Well all of my Einstein WUs for the Global Correlations S5 Engineering app exit with a signal 11 compute error. I cannot deselect this app from the preferences. see here:

http://einsteinathome.org/task/173486450
http://einsteinathome.org/task/173486469
http://einsteinathome.org/task/173485657

Ageless
Joined: 26 Jan 05
Posts: 2,949
Credit: 5,374,792
RAC: 0

Global Correlations S5 - Signal 11

Is your Linux a 64 bit variety? If so, did you load its 32bit compatibility libraries? If you're not sure, check this BOINC wiki entry for more information.

Einstein@Home doesn't have any 64bit applications, only 32bit versions, which not all Linux distros know what to do with automatically.

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

Looks like that could be it.

Looks like that could be it. I swore I had all the 32bit compat libs installed, because I remember having to do that for other projects. Is there no way to disable that application from my preferences? The other apps seem to crunch fine.

Ageless
Joined: 26 Jan 05
Posts: 2,949
Credit: 5,374,792
RAC: 0

No, ABP is the optional one.

Message 97934 in response to message 97933

No, ABP is the optional one. The S5xxx searches are the normal ones.

Bikeman wrote here: "This has to do with the main scientific focus of the project (LIGO data processing), and there are also constraints I guess because of the grants that E@H received, e.g. from NSF. If you get grants for growing oranges, you can grow apples with spare ressources, but you shouldn't be able to switch completly to growing apples...at the end you have to show some nice oranges to the funding bodies :-)"

I can follow that perfectly.

Mind, there are other explanations for signal 11. See this FAQ which gives more options.

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

Hmm, I checked my

Hmm, I checked my system:

ldd einstein_S5GCE_1.04_i686-pc-linux-gnu__S5GCESSE2
linux-gate.so.1 => (0xffffe000)
libdl.so.2 => /lib32/libdl.so.2 (0xf7f38000)
libpthread.so.0 => /lib32/libpthread.so.0 (0xf7f20000)
libm.so.6 => /lib32/libm.so.6 (0xf7efa000)
libc.so.6 => /lib32/libc.so.6 (0xf7db7000)
/lib/ld-linux.so.2 (0xf7f47000)

ldd einsteinbinary_ABP2_1.08_i686-pc-linux-gnu
linux-gate.so.1 => (0xffffe000)
libpthread.so.0 => /lib32/libpthread.so.0 (0xf7f30000)
libm.so.6 => /lib32/libm.so.6 (0xf7f0a000)
libc.so.6 => /lib32/libc.so.6 (0xf7dc7000)
/lib/ld-linux.so.2 (0xf7f53000)

The required 32bit libraries are installed....

I did find a different error as I was going through them.... here:

http://einsteinathome.org/task/173761592

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

I suspect the signal 11

I suspect the signal 11 errors were cause by some file system errors that I found. Now I seem to be getting much higher ratio of success vs error WUs and the errors that I do find now appear to be that new type exit(13).

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 3,904
Credit: 189,164,559
RAC: 24,739

There actually is a problem

There actually is a problem with GCE/C1 Linux Apps that causes a signal 11 (segfault). I started looking into this but was distracted by other more urgent things. Might well be a bug left in the App; I could use some help.

BM

BM

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 3,904
Credit: 189,164,559
RAC: 24,739

If you experience this

If you experience this problem, please do the following:

Put a file named "EAH_DEBUG_GDB" (w/o double quotes) in the BOINC/ directory that contains the following three lines:

cont
bt
quit

Whenever an Einstein@Home S5GC1 Application is started then, a gdb should attach to it and wait for something to happen (e.g. a signal or a normal termination of the App). If this happens, the gdb(!) should write a stackdump into stderr.txt (that is reported back to the server) and then quit.

An attached gdb shouldn't hurt performance too much.

BM

BM

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

I will have to check


I will have to check tomorrow, I am not sure if I have GDB installed. If I do I will see if I can get you some stack dumps.

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

Ok, I did not have GDB

Ok, I did not have GDB installed, I do now on one host and it is crunching away to see if/when I get some crashes.

here is the host:

http://einsteinathome.org/host/2772939

Mr. Hankey
Mr. Hankey
Joined: 30 Apr 10
Posts: 9
Credit: 103,466,998
RAC: 0

Just a quick update.. The

Just a quick update.. The host is happily churning out completed WUs since installing GDB and trying to catch a sig 11. I think though that I may not be getting the WUs that were triggering the issue before. I am getting the S5 #1 WUs and I recall the issue was with the plain S5 Wus. I will let it carry on....

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.