Re: [squid-users] squid 3.2.1 under solaris dying with "segment violation"

Submit a bug in bugzilla, please. http://bugs.squid-cache.org/

On Mon, Sep 17, 2012 at 8:54 AM, Jose-Marcio Martins da Cruz
<Jose-Marcio.Martins_at_mines-paristech.fr> wrote:
>
> Hello,
>
> Someone can help me ?
>
> Shall I send this to bugs_at_squid-cache.org ?
>
> As from the cores I posted here, squid dies each 1-3 minutes some place at
> memory allocation/free place, given some bad ipcache entry.
>
> squid stop dying friday evening and start dying again monday morning.
> Probably some client asking for some "strange URL". This doesn't happens
> with squid 3.1.19.
>
> Is there any other information I shall post ?
>
> Thanks
>
> José-Marcio
>
>
> Jose-Marcio Martins da Cruz wrote:
>>
>>
>> Hello,
>>
>> No answers about this problem... I finally succeeded to get some core
>> files.
>>
>> I have some number of core files, all of them can be resumed by these two,
>> related to ipcache management : ipcacheFreeEntry or ipcacheCreateEntry.
>>
>> I also got some results with truss, showing that but crash are related to
>> ipcache entries.
>>
>> Any help will be welcome
>>
>> Regards,
>>
>> José-Marcio
>>
>> ***********************************************************************
>> ...
>> Core was generated by `(squid-1)'.
>> Program terminated with signal 6, Aborted.
>> #0 0xfed59315 in _lwp_kill () from /lib/libc.so.1
>> (gdb) bt
>> #0 0xfed59315 in _lwp_kill () from /lib/libc.so.1
>> #1 0xfed54188 in thr_kill () from /lib/libc.so.1
>> #2 0xfed01d73 in raise () from /lib/libc.so.1
>> #3 0xfece1bbd in abort () from /lib/libc.so.1
>> #4 0x082288a1 in death (sig=11) at tools.cc:388
>> #5 0xfed55dbf in __sighndlr () from /lib/libc.so.1
>> #6 0xfed4bab5 in call_user_handler () from /lib/libc.so.1
>> #7 <signal handler called>
>> #8 0x081dcb3b in memFree (p=0x87f4af8, type=184436712) at mem.cc:214
>> #9 0x081d6bd0 in ipcacheFreeEntry (data=0xafe47e8) at ipcache.cc:1140
>> #10 0x081d7342 in ipcacheRelease (dofree=true, i=0xafe47e8) at
>> ipcache.cc:203
>> #11 ipcacheRelease (i=0xafe47e8, dofree=true) at ipcache.cc:186
>> #12 0x081d8aba in ipcache_nbgethostbyname (name=0xbdd4440 <error reading
>> variable>,
>> handler=0x81f34a0 <peerSelectDnsResults(ipcache_addrs const*,
>> DnsLookupDetails const&, void*)>, handlerData=0xa841940)
>> at ipcache.cc:670
>> #13 0x081f0dd8 in peerSelectDnsPaths (psstate=0xa841940) at
>> peer_select.cc:268
>> #14 0x081f1a1d in peerSelectFoo (ps=0xa841940) at peer_select.cc:511
>> #15 0x0822e324 in tunnelStart (http=0xbdd1e38, size_ptr=0xbdd1e50,
>> status_ptr=0xbdd1ef4) at tunnel.cc:665
>> #16 0x08161bea in ClientHttpRequest::processRequest (this=0xbdd1e38) at
>> client_side_request.cc:1340
>> #17 0x081623dd in ClientHttpRequest::doCallouts (this=0xbdd1e38) at
>> client_side_request.cc:1604
>> #18 0x08164545 in ClientRequestContext::clientAccessCheckDone
>> (this=0xa5d1b48,
>> answer=@0x80469cc: ACCESS_ALLOWED)
>> at client_side_request.cc:842
>> ...
>>
>> *********************************************************
>>
>> Core was generated by `(squid-1)'.
>> Program terminated with signal 6, Aborted.
>> #0 0xfed59315 in _lwp_kill () from /lib/libc.so.1
>> (gdb) bt
>> #0 0xfed59315 in _lwp_kill () from /lib/libc.so.1
>> #1 0xfed54188 in thr_kill () from /lib/libc.so.1
>> #2 0xfed01d73 in raise () from /lib/libc.so.1
>> #3 0xfece1bbd in abort () from /lib/libc.so.1
>> #4 0x082288a1 in death (sig=11) at tools.cc:388
>> #5 0xfed55dbf in __sighndlr () from /lib/libc.so.1
>> #6 0xfed4bab5 in call_user_handler () from /lib/libc.so.1
>> #7 <signal handler called>
>> #8 0x081dcaca in memAllocate (type=142559992) at mem.cc:206
>> #9 0x081d6bfa in ipcacheCreateEntry (name=0xc042a10 <error reading
>> variable>)
>> at ipcache.cc:302
>> #10 0x081d8adf in ipcache_nbgethostbyname (name=0xc042a10 <error reading
>> variable>,
>> handler=0x81f34a0 <peerSelectDnsResults(ipcache_addrs const*,
>> DnsLookupDetails const&, void*)>, handlerData=0xa5d2de0)
>> at ipcache.cc:692
>> #11 0x081f0dd8 in peerSelectDnsPaths (psstate=0xa5d2de0) at
>> peer_select.cc:268
>> #12 0x081f1a1d in peerSelectFoo (ps=0xa5d2de0) at peer_select.cc:511
>> #13 0x0818b83f in FwdState::start (this=0xa6061d0, aSelf=...) at
>> forward.cc:144
>> #14 0x0818fa4d in FwdState::Start (clientConn=..., entry=0xc08f0e8,
>> request=0xc042930, al=...) at forward.cc:318
>> #15 0x0815c3a1 in clientReplyContext::processMiss (this=0xc041768) at
>> client_side_reply.cc:665
>> #16 0x0815cab0 in clientReplyContext::doGetMoreData (this=0xc041768) at
>> client_side_reply.cc:1754
>> #17 0x0815cf6b in clientReplyContext::identifyStoreObject (this=0xc041768)
>> at
>> client_side_reply.cc:1535
>> #18 0x08160974 in ClientHttpRequest::httpStart (this=0xc040308) at
>> client_side_request.cc:1358
>> #19 0x081623dd in ClientHttpRequest::doCallouts (this=0xc040308) at
>> client_side_request.cc:1604
>> #20 0x081635d5 in checkNoCacheDoneWrapper (answer=ACCESS_ALLOWED,
>> data=0xa5d1b48) at client_side_request.cc:1271
>> ---Type <return> to continue, or q <return> to quit---
>> #21 0x08284655 in ACLChecklist::checkCallback (this=0xa5d25a0,
>> answer=ACCESS_ALLOWED) at Checklist.cc:194
>> #22 0x082514b3 in ACLFilledChecklist::checkCallback (this=0xa5d25a0,
>> answer=ACCESS_ALLOWED) at FilledChecklist.cc:37
>> #23 0x08285403 in ACLChecklist::matchNonBlocking (this=0xa5d25a0) at
>> Checklist.cc:126
>> #24 0x08285739 in ACLChecklist::nonBlockingCheck (this=0xa5d25a0,
>> callback_=0x81635b0 <checkNoCacheDoneWrapper(allow_t, void*)>,
>> callback_data_=0xa5d1b48) at Checklist.cc:325
>> #25 0x08163589 in ClientRequestContext::checkNoCache (this=0xa5d1b48) at
>> client_side_request.cc:1256
>> #26 0x08162371 in ClientHttpRequest::doCallouts (this=0xc040308) at
>> client_side_request.cc:1557
>> #27 0x08164545 in ClientRequestContext::clientAccessCheckDone
>> (this=0xa5d1b48,
>> answer=@0x80469cc: ACCESS_ALLOWED)
>> at client_side_request.cc:842
>> #28 0x081647b9 in ClientRequestContext::clientAccessCheck2
>> (this=0xa5d1b48) at
>> client_side_request.cc:735
>> #29 0x08161e35 in ClientHttpRequest::doCallouts (this=0xc040308) at
>> client_side_request.cc:1542
>> #30 0x08164545 in ClientRequestContext::clientAccessCheckDone
>> (this=0xa5d1b48,
>> answer=@0x8046d70: ACCESS_ALLOWED)
>> at client_side_request.cc:842
>> #31 0x081648e5 in clientAccessCheckDoneWrapper (answer=ACCESS_ALLOWED,
>> data=0xa5d1b48) at client_side_request.cc:747
>> #32 0x08284655 in ACLChecklist::checkCallback (this=0xa5d2790,
>> answer=ACCESS_ALLOWED) at Checklist.cc:194
>> #33 0x082514b3 in ACLFilledChecklist::checkCallback (this=0xa5d2790,
>> answer=ACCESS_ALLOWED) at FilledChecklist.cc:37
>> #34 0x082854b1 in ACLChecklist::matchNonBlocking (this=0xa5d2790) at
>> Checklist.cc:105
>> #35 0x08285739 in ACLChecklist::nonBlockingCheck (this=0xa5d2790,
>> callback_=0x81648c0 <clientAccessCheckDoneWrapper(allow_t, void*)>,
>> callback_data_=0xa5d1b48) at Checklist.cc:325
>> #36 0x08164a17 in ClientRequestContext::clientAccessCheck (this=0xa5d1b48)
>> at
>> client_side_request.cc:715
>> #37 0x08161cc9 in ClientHttpRequest::doCallouts (this=0xc040308) at
>> client_side_request.cc:1513
>> #38 0x08165ba9 in ClientRequestContext::hostHeaderVerify (this=0xa5d1b48)
>> at
>> client_side_request.cc:663
>> #39 0x08161c97 in ClientHttpRequest::doCallouts (this=0xc040308) at
>> client_side_request.cc:1506
>> ---Type <return> to continue, or q <return> to quit---
>> #40 0x08151348 in clientProcessRequest (conn=0xb88e568, hp=0xc042900,
>> context=<optimized out>, method=..., http_ver=...)
>> at client_side.cc:2691
>> #41 0x08152329 in ConnStateData::clientParseRequests (this=0xb88e568) at
>> client_side.cc:2788
>>
>>
>>
>> Jose-Marcio Martins da Cruz wrote:
>>>
>>>
>>> Hello,
>>>
>>> I'm begining a new thread.
>>>
>>> To remember, I'm trying to migrate a squid server from 3.1.19 to 3.2.1.
>>> The old
>>> one works fine, but with the new version, squid dies each 1 to 3 minutes,
>>> with
>>> "segment violation".
>>>
>>> I'm not able do get core files under Solaris, despite of correct settings
>>> (ulimit, ...).
>>>
>>> So, I set up debug_options to "ALL,8" and "ALL,9". and I wrote a single
>>> script
>>> which prints the last 25 lines before each SEGFAULT received.
>>>
>>> The result is at http://www.j-chkmail.org/users/squid
>>>
>>> The contents of the three files are :
>>>
>>> * cores-4.txt - debug_options ALL,8 - squid doing cache
>>> * cores-5.txt - debug_options ALL,9 - squid doing cache
>>> * cores-9.txt - debug_options ALL,8 - cache_dir option commented
>>>
>>> (If you want, I can send you the files.)
>>>
>>> The third test was to be sure that the problem didn't come from the cache
>>> partition on a zfs filesystem.
>>>
>>>
>>> Before each "segfault" there is a line indicating the destination
>>> hostname. Most
>>> of the time, but not allways, the hostname doesn't resolve or it's a
>>> CNAME.
>>>
>>> Thanks for your help,
>>>
>>> José-Marcio
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>
>
> --
>
> Envoyé de ma machine à écrire.
> ---------------------------------------------------------------
> Spam : Classement statistique de messages électroniques -
> Une approche pragmatique
> Chez Amazon.fr : http://amzn.to/LEscRu ou http://bit.ly/SpamJM
> ---------------------------------------------------------------
> Jose Marcio MARTINS DA CRUZ http://www.j-chkmail.org
> Ecole des Mines de Paris http://bit.ly/SpamJM
> 60, bd Saint Michel 75272 - PARIS CEDEX 06
> mailto:Jose-Marcio.Martins_at_mines-paristech.fr

-- 
    /kinkie