Segmentation fault...

K

knopflerbruce

Limp Gawd
Joined
Nov 8, 2007
Messages
172
I keep getting a segmentation fault on a fresh 10.10 install on a 2P G34 setup. Never seen it before, and i have no clue on how to fix it. This is what the terminal says:

"Downloading new core: FahCore_a5.exe
Segmentation fault" - and then fah quits.
 
Are you running a 32 or 64 bit system?

Are you running FAHv6 or v7?
Do you have FAH installed into a home user directory?
Does FAH have write permissions to the directory?
 
64bit, 6.34.

It's in the home directory I think ( ~/fah in terminal).

How do I check the write permissions?
 
I set fah6 to 777 through chmod now, it reads -rwxrwxrwx now, did the same thing to Core_a5.fah. Still the same error, though.
 
I would never have guessed that...

musky is on top of his game.

Odd, since I run a swapless system here without any issue
 
Straaaange (same here). What .iso did you use to install?
Did you choose any extra options during the install?
(or was it just 'click next until finished'?)

Can you do:
Code: 
sudo -s
echo 1 > /proc/sys/kernel/print-fatal-signals
exit

Then rerun the client and:
1. Capture complete terminal output
2. Capture crash part of 'dmesg'
 
ubuntu-10.10-desktop-amd64.iso

I don't think I changed anything, no. Was pretty straight forward. Didn't even download updates while installing. Did that part when the setup was complete - so yes, pretty much 'next' all the way through.

I tried what you just said, but the dmesg output I get is daaaaaaamn long. Is this what you're looking for?

"[ 31.523878] fah6[1870]:segfault at 4 ip 000000000041e7ba sp 00007f68cb520a10 error 4 in fah6[400000+49000]
[ 2763.310733] fah6[1941]: segfault at 4 ip 000000000041e7ba sp 00007fd5901e0a10 error 4 in fah6[400000+49000]
[ 2763.410760] fah6/1941: potentially unexpected fatal signal 11."

The output when launching fah looks like this:

Launch directory: /home/knopflerbruce/fah
Executable ./fah6
Arguments: -verbosity 9 -smp 24 -bigadv

Ask before connecting: no
User name: knopflerbruce (Team 37651)
User ID: .................
Machine ID: 1

Loaded queue successfully

+ processing work unit
core required. FahCore-a5.exe
Core not found
- Core is not present or corrupted.
- Attempting to download new core...
+ Downloading new core. FahCore_a5.exe
Downloading core (/~pande/Linux/AMD64/Core-a5.fah from www.stanford.edu)
Segmentation fault
 
Gotcha, thanks.

So... if I was to reproduce it. I should install the OS and do a system update post-install, correct?

It definitely looks like name resolution issue (weird, shouldn't be the case with dynamically
linked binaries). So you're saying that you tried enabling nscd to no avail, is that true?

Thanks,
tear
 
Yes, I tried that (changed no to yes somewhere, as described earlier). Feel free to give it a shot :) I do have the 12.10 release here, maybe I should try that tomorrow... (4AM here, so can't stay up much longer).

12.10 worked flawlessly on my single-CPU systems at least.
 
That's the thing, I do have 10.10 in a VM but haven't had a problem -- I'll start from scratch.

Hmm, nscd should be a separate package... Does:
Code: 
sudo dpkg -l | grep -i nscd
return anything?
 
and
Code: 
ps auxw | grep [^]]nscd
gives you something back as well?
 
So it is running.... let's kill it and see if THAT makes a difference :)

killall -11 nscd

and then try fah again....

If it doesn't work, I'll start digging deeper at my end :)
 
Are you sure the system is stable? Seg faults happen a lot with bad hardware. Might be something to look into.
 
Nope, same error - had to add sudo to be permitted to kill it, though... (the ps auxw thingy doesn't give any output, so I guess nscd isn't running anymore).
 
I can't be 100% sure, but at the moment the system is running at stock (1.6GHz), and I was benchmarking it at 3.4 in server 2k8 just before installing 10.10. What about AHCI/IDE settings? Irrelevant? I set it to IDE (using an old 32gb SSD).
 
Shouldn't...

I'm outta ideas; will go about installing 10.10 from the same .iso + all updates...
 
cool, thanks alot! I will log in again in about 7h, if you still haven't figured it out by then I'll just give the other image a shot, the one I used for my 3930k rigs a while back - with no issues. Hopefully kraken works...
 
One more Q if you're still around... what are the extra command-line params in your client.cfg and are you adding any in the command line?
 
Borkage appears to be somewhere in client's IPv6 code path.

I was able to reproduce the problem by enabling IPv6... that said, disable IPv6 and that should get you going :)

Code: 
[06:14:44] Loaded queue successfully.
[06:14:44]
[06:14:44] - Autosending finished units... [December 18 06:14:44 UTC]
[06:14:44] + Processing work unit
[06:14:44] Trying to send all finished work units
[06:14:44] Core required: FahCore_a5.exe
[06:14:44] + No unsent completed units remaining.
[06:14:44] Core not found.
[06:14:44] - Autosend completed
[06:14:44] - Core is not present or corrupted.
[06:14:44] - Attempting to download new core...
[06:14:44] + Downloading new core: FahCore_a5.exe
[06:14:44] Downloading core (/~pande/Linux/AMD64/beta/Core_a5.fah from www.stanford.edu)
Segmentation fault
fah@fah-VirtualBox:~/fah-6.34$ dmesg | tail -22
[ 1428.541121] fah6[2098]: segfault at 4 ip 000000000041e7ba sp 00007f49643d0a10 error 4 in fah6[400000+49000]
[ 1428.541146] fah6/2098: potentially unexpected fatal signal 11.
[ 1428.541157]
[ 1428.541167] CPU 0
[ 1428.541177] Modules linked in: binfmt_misc ppdev parport_pc i2c_piix4 psmouse serio_raw lp parport floppy e1000
[ 1428.541195]
[ 1428.541206] Pid: 2098, comm: fah6 Not tainted 2.6.35-22-generic #33-Ubuntu /VirtualBox
[ 1428.541216] RIP: 0033:[<000000000041e7ba>]  [<000000000041e7ba>] 0x41e7ba
[ 1428.541228] RSP: 002b:00007f49643d0a10  EFLAGS: 00010202
[ 1428.541237] RAX: 0000000000000000 RBX: 00007f49643d0a20 RCX: 0000000000000192
[ 1428.541245] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 00007f49643d0a3c
[ 1428.541253] RBP: 00007f49643d3390 R08: 000000000000001c R09: 0101010101010101
[ 1428.541261] R10: 00007f49643d0790 R11: 00007f496445a85e R12: 00007f49643d380c
[ 1428.541268] R13: 0000000000000055 R14: 0000000000000000 R15: 0000000000000005
[ 1428.541282] FS:  00007f49643d4710(0000) GS:ffff880001e00000(0000) knlGS:0000000000000000
[ 1428.541290] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1428.541297] CR2: 0000000000000004 CR3: 000000002f806000 CR4: 00000000000006f0
[ 1428.541305] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1428.541312] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 1428.541319] Process fah6 (pid: 2098, threadinfo ffff88003b96a000, task ffff88003cf596e0)
[ 1428.541325]
[ 1428.541330] Call Trace:
 
Last edited:
It worked! didn't have enough time to login here before i had to leave, but I got the client running. Not sure if my frame times are as good as I expected, but I'll investigate that part tonight. 24c 1.6Ghz isn't really THAT much power... if I double the frequency... THEN! Hehe.
 
You must log in or register to reply here.
Back
Top