and why? i know it's not the nicest stack around but it's the only one i have so please don't smash it

doggybob@bone1:~/boinc$ ./boinc
05-Dec-2012 11:29:52 [---] Starting BOINC client version 6.12.34 for x86_64-pc-linux-gnu
05-Dec-2012 11:29:52 [---] log flags: file_xfer, sched_ops, task
05-Dec-2012 11:29:52 [---] Libraries: libcurl/7.21.0 OpenSSL/0.9.8o zlib/1.2.3.4 libidn/1.15 libssh2/1.2.6
05-Dec-2012 11:29:52 [---] Data directory: /home/doggybob/boinc
05-Dec-2012 11:29:52 [---] Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
05-Dec-2012 11:29:52 [---] Processor: 8.00 MB cache
05-Dec-2012 11:29:52 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor
05-Dec-2012 11:29:52 [---] OS: Linux: 2.6.32-5-amd64
05-Dec-2012 11:29:52 [---] Memory: 23.58 GB physical, 27.94 GB virtual
05-Dec-2012 11:29:52 [---] Disk: 641.84 GB total, 408.43 GB free
05-Dec-2012 11:29:52 [---] Local time is UTC +1 hours
05-Dec-2012 11:29:52 [---] No usable GPUs found
05-Dec-2012 11:29:52 [---] A new version of BOINC is available. <a href=htt////boinc.berkeley.edu/download.php>Download it.<>a>
05-Dec-2012 11:29:52 [Asteroids@home] URL http://asteroidsathome.net/boinc/; Computer ID 3469; resource share 10
05-Dec-2012 11:29:52 [NRG] URL http://boinc.med.usherbrooke.ca/nrg/; Computer ID 4328; resource share 10
05-Dec-2012 11:29:52 [climateprediction.net] URL http://climateprediction.net/; Computer ID 1252571; resource share 10
05-Dec-2012 11:29:52 [OProject@Home] URL http://oproject.info/; Computer ID 4656; resource share 10
05-Dec-2012 11:29:52 [Radioactive@Home] URL http://radioactiveathome.org/boinc/; Computer ID 4727; resource share 10
05-Dec-2012 11:29:52 [NumberFields@home] URL http://NumberFields.asu.edu/NumberFields/; Computer ID 8865; resource share 10
05-Dec-2012 11:29:52 [---] General prefs: from http://gerasim.boinc.ru/ (last modified 24-Jul-2011 08:52:45)
05-Dec-2012 11:29:52 [---] Host location: none
05-Dec-2012 11:29:52 [---] General prefs: using your defaults
05-Dec-2012 11:29:52 [---] Reading preferences override file
05-Dec-2012 11:29:52 [---] Preferences:
05-Dec-2012 11:29:52 [---] max memory usage when active: 21729.33MB
05-Dec-2012 11:29:52 [---] max memory usage when idle: 23902.27MB
05-Dec-2012 11:29:52 [---] max disk usage: 10.00GB
05-Dec-2012 11:29:52 [---] suspend work if non-BOINC CPU load exceeds 80 %
05-Dec-2012 11:29:52 [---] (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
05-Dec-2012 11:29:52 [---] Not using a proxy
Initialization completed
05-Dec-2012 11:29:52 [OProject@Home] Restarting task final-sp-14464_0 using shor-sp version 104
05-Dec-2012 11:29:52 [Radioactive@Home] Restarting task sample_1197814_0 using radac version 169
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_264_1 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_283_1 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_126_0 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_290_1 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_289_1 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_121_271_0 using period_search version 10100
05-Dec-2012 11:29:52 [Asteroids@home] Restarting task ps_121130_122_38_1 using period_search version 10100
*** stack smashing detected ***: ./boinc terminated
SIGSEGV: segmentation violation
Segmentation fault

i've restarted the client 5 times and it just keeps crashing with the segmentation fault

i tried starting the client with the --start-delay parm set to 120 secs to delay starting the science applications for 2 minutes in case one of those is smashing my lovely stack but it makes no difference, the client crashes almost immediately

if it isn't one of the science apps then it must be BOINC client itself, googled "stack smashing" and found interesting stuff from which I quote...

Stack Smashing is actually a protection mechanism used by gcc to detect buffer overflow attacks.

...which makes me suspicious that BOINC is a virus trying to attack my system.

of course the first thing to try when crap happens is a reboot but the machine is in the middle of a big job I dare not interrupt, have a few other NCI apps running and they're not having any problem so it seems limited to BOINC client which i won't start again until after i reboot or i hear some advice/comments/drivel from whomever/whatever

will google around a bit more, seems others have seen "stack smashing" outside of boincville so gotta be some advice somewhere

---

"Windows" -- an American English word, meaning "A real operating system is too hard for me."

I'll send it off to the developers, but I know already what they're going to ask: Can you reproduce this effect on BOINC 7.0? If not, it's already fixed in that one. But if you can, they'll be very interested.

---

i searched all of boincdom (trac, this forum, boinc faq, boinc wiki, wikipedia articles on boinc and of course google, bing and other search engines for "stack smashing" related to boinc, "stak smashing" and several other likely misspellings and turned up nothing therefore I conclude the boinc developers have never heard of it in connection with boinc, possibly have never heard of it in any context other than "my stack of wine glasses fell over and got smashed" and have therefore made no conscious effort to fix it, so how can it possibly be fixed in 7.0 except by sheer luck? that is my pre-response to their response

one of the projects I crunch requires boinc 6.12 and will not issue tasks to 7.* clients, of course that doesn't mean I can't try 7.* but i'm not crazy about it, i'll wait see what happens after a reboot maybe it'll just go away

---

"Windows" -- an American English word, meaning "A real operating system is too hard for me."

this may not be a problem with boinc after all, i mentioned another big job i have running and it turns out that one is using 20GB of the 24GB RAM in the machine, i believe it should be using no more than 4GB but not 100% sure

could be that application is causing the problem and boinc is merely tripping over the mounds of rubble it's leaving behind... a victim rather than the perpetrator

i'll get back when i have more info

---

"Windows" -- an American English word, meaning "A real operating system is too hard for me."

one of the projects I crunch requires boinc 6.12 and will not issue tasks to 7.* clients

Which one may that be and what is their reasoning behind it?

---

one of the projects I crunch requires boinc 6.12 and will not issue tasks to 7.* clients

Which one may that be and what is their reasoning behind it?

it was oproject and I say "was" because it seems they updated/changed server code yesterday and now 7.* clients work but it's all very confusing because one other volunteer claims he has been running 7.* clients with oproject for several weeks, on the other hand the admin claimed (up until yesterday) that 7.* would not work, now after the apparent server code update/change/whatever he claims he isn't sure which clients will and will not work! oh, and one volunteer reported (before the server update/change/whatever) that her 7.0.31 beta client received a message from oproject server saying her client is too old!!

oproject admin's English is not very good and it seems he either has difficulty understanding the server docs or has difficulty explaining to the volunteers how things are configured to work but he trys hard so i stay

on top of all that fun and confusion is the fact that the latest boinc package from the distro I run is 6.12.28, the 7.* versions from the boinc site do not run on my distro due to absent shared libs, yes my OS is up to date, so it seems the only way i can try a 7.* is to compile my own which so far has led to the same absent shard libs issues i encounter when attempting to run the standard 7.* client, i know there is likely a reasonable solution to this version and shared libs problem but I haven't been sufficiently motivated to find that solution so far as 6.12.28 has been working fine for me

the deeper i dig the more i believe this stack smashing thing is not really boinc's fault it's some other misbehaving app doing it and boinc is merely detecting and reporting the problem or something like that

---

"Windows" -- an American English word, meaning "A real operating system is too hard for me."

solved and fixed!

it turns out it was not the app i mentioned earlier in this thread, the one using 89% of the host's RAM, that's a problem but after the job it was running completed and i was able to shutdown that app the stack smashing problem reported by boinc persisted so it must no be the dastardly stack smasher

the real smasher was hinted at in the URL i mentioned earlier where they said the stack smashing error can indicate the app is trying to read a 20 (for example) char input into a 19 char buffer, i hypothesized that boinc was trying to read from one of the config or project related files in the boinc data dir and failing due to a too small input buffer

to test that hypothesis i decided to delete the files one-by-one to see if the sack smash error would go away, i ran the boinc command with the --detach_project URL parm on each project followed by the the boinc command with no parms, after the second detach the boinc command did not crash with the stack smashing error so i conclude that one of the data files associated with that project had corrupted and boinc was trying to read input from that file into a buffer that was too small

after that i re-attached to all the projects i detached from and am now back to normal

thank you watson, it was a right bit of detective work, detoured for a while by that RAM hogging app but we nailed the bugger in the end

sherlock

---

"Windows" -- an American English word, meaning "A real operating system is too hard for me."