[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ldm-5.2 problems



On Fri, 9 Aug 2002, [ISO-8859-1] Christian Pagé wrote:

> Hi,
> 
> Here is the dbx info:
> 
> 312 [/io/ldm] % dbx runtime/bin/pqact
> dbx version 7.3 MR 55458_Apr30_MR Apr 30 1999 13:44:41
> Core from signal SIGBUS: Bus error
> (dbx) t
>  >  0 __malloc(0x0, 0x0, 0x4, 0x0, 0x88, 0x0, 0xfffffffc, 0x10039e7c) 
> ["/xlv44/6.5.13m/work/irix/lib/libc/libc_n32_M4/gen/malloc.c":209, 
> 0xfb0ba94]
>     1 _malloc(0x0, 0x0, 0x4, 0x0, 0x88, 0x0, 0xfffffffc, 0x10039e7c) 
> ["/xlv44/6.5.13m/work/irix/lib/libc/libc_n32_M4/gen/malloc.c":186, 
> 0xfb0ba28]
>     2 lmatcher(0x10039eb8, 0x0, 0x5, 0x1003a178, 0x88, 0x1, 0xfffffffc, 
> 0x10039e7c) 
> ["/xlv44/6.5.13m/work/irix/lib/libc/libc_n32_M4/gen/engine.c":249, 
> 0xfa4cb28]
>     3 __nregexec(0x90, 0x0, 0x4, 0x0, 0x88, 0x0, 0x10039eb8, 0x10039e7c) 
> ["/xlv44/6.5.13m/work/irix/lib/libc/libc_n32_M4/gen/regexec.c":275, 
> 0xfa4f7bc]

Christian,

Good news for me, the crash is not because of the surf_split.c fix I made.
I believe it's  with the regular expression package on your machine or a
pqsurf.conf  entry.  Here's how I would test it.  First comment out all
entries in the etc/pqsurf.conf file. Start the ldm, one can run pqsurf
from the command line so not to crash the whole ldm. Start pqsurf like
this:

% pqsurf -vl -

At this point it should run ok.  Stop/Start pqsurf adding in one more
pqsurf entry until you find the culprit. If it's a PIPE entry culprit, try
running it from the command line also.  

The actual line in the code that's failing in palt.c is:

                if((infop->feedtype & pal->feedtype)
                   && ((regexec(&pal->prog, infop->ident,
                                pal->prog.re_nsub +1,  pal->pmatchp,
                                        0) == 0)
                       || (strcmp(pal->pattern, "^_ELSE_$") == 0
                           && did_something == 0
                           && infop->ident[0] != '_')))

According to the dbx output, it's the (regexec(... part that's failing.

 
Let me know what you find out, so we can track this down.

Robb...




>     4 processProduct(0x7fff23f8, 0x0, 0x4016ab8, 0xa0, 0x88, 0x0, 
> 0xfffffffc, 0x10039e7c) ["/io/ldm/ldm-5.2/src/pqact/palt.c":779, 
> 0x100069c4]
>     5 pq_sequence(0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0xfffffffc, 0x10039e7c) 
> ["/io/ldm/ldm-5.2/src/pq/pq.c":5147, 0x1001649c]
>     6 main(0xb, 0x100068e0, 0x7fffffff, 0xf423f, 0x1, 0x0, 0xfffffffc, 
> 0xffffffff) ["/io/ldm/ldm-5.2/src/pqact/pqact.c":439, 0x10004ccc]
>     7 __start() ["/xlv55/kudzu-
> apr12/work/irix/lib/libc/libc_n32_M4/csu/crt1text.s":177, 0x10004298]
> (dbx) quit
> 
> > The signal 10 dumps a core file, so could you run dbx on the file. Do
> > % file core
> > to see if it's pqsurf core. I think the syntax is
> >
> > % dbx pqsurf core
> >
> > This should show the problem.  Otherwise, if I could get the core file 
> > to
> > do some debugging.  Is this a 32 bit build? Any other info would 
> > help,logs
> > Also, are you sure that the pqsurf file was replace?  I'm on vacation
> > tomorrow, so I'll look at it Monday.   If you need, you can us the 
> > pqusrf
> > from the LDM 5.1.4 distribution with the LDM 5.2 release.
> 
> This is an -n32 build, so 32-bit. I tried using pqsurf from ldm-5.1.4 
> but got the same problems, but with signal 11.
> 
> Here are the logs:
> 
> Aug  9 13:26:01 6Q:io ldm: LDM Start (login name, tty, login time, 
> [from]): ldm        ttyq37       Aug  8 17:44  (132.208.133
> .37)
> Aug 09 13:26:01 5Q:io rpc.ldmd[77312290]: Starting Up (built: Jul 30 
> 2002 14:47:27)
> Aug 09 13:26:01 5Q:io pqbinstats[77339594]: Starting Up (77312290)
> Aug 09 13:26:01 5Q:io pqact[77024855]: Starting Up
> Aug 09 13:26:01 5Q:io pqsurf[77304103]: Starting Up (77312290)
> Aug 09 13:26:01 5Q:io 128.174.80.47[71288800]: run_requester: Starting 
> Up: 128.174.80.47
> Aug 09 13:26:01 5Q:io pqact[77843305]: Starting Up
> Aug 09 13:26:01 5Q:io 128.174.80.47[71288800]: run_requester: 
> 20020809132548.723 TS_ENDT {{IDS|DDPLUS,  ".*"}}
> Aug 09 13:26:01 5Q:io flood[77628265]: run_requester: Starting Up: 
> flood.atmos.uiuc.edu
> Aug 09 13:26:01 5Q:io dns2[78016494]: run_requester: Starting Up: 
> dns2.cmc.ec.gc.ca
> Aug 09 13:26:01 5Q:io flood[77628265]: run_requester: 20020809132523.450 
> TS_ENDT {{UNIWISC|HDS,  ".*"}}
> Aug 09 13:26:01 5Q:io striker[77959463]: run_requester: Starting Up: 
> striker.atmos.albany.edu
> Aug 09 13:26:01 5Q:io 128.174.80.47[71288800]: FEEDME(128.174.80.47): OK
> Aug 09 13:26:01 5Q:io flood[77628265]: FEEDME(flood.atmos.uiuc.edu): OK
> Aug 09 13:26:01 5Q:io striker[77959463]: run_requester: 
> 20020809132435.257 TS_ENDT {{NLDN,  ".*"}}
> Aug 09 13:26:01 5Q:io striker[77959463]: 
> FEEDME(striker.atmos.albany.edu): OK
> Aug 09 13:26:02 5Q:io dns2[78016494]: run_requester: 20020809122601.130 
> TS_ENDT {{GEM,  ".*"}}
> Aug 09 13:26:02 5Q:io dns2[78016494]: FEEDME(dns2.cmc.ec.gc.ca): OK
> Aug 09 13:26:18 5Q:io pqsurf[77304103]: child 77843305 terminated by 
> signal 10
> Aug 09 13:26:18 5Q:io pqsurf[77304103]: Exiting
> Aug 09 13:26:18 5Q:io pqsurf[77304103]:   Queue usage (bytes): 2015232
> Aug 09 13:26:18 5Q:io pqsurf[77304103]:            (nregions):   11470
> Aug 09 13:26:18 5Q:io pqsurf[77304103]: Number of products 29
> Aug 09 13:26:18 5Q:io pqsurf[77304103]: Number of observations 81
> Aug 09 13:26:18 5Q:io pqsurf[77304103]: Number of dups 5
> Aug 09 13:26:24 5Q:io rpc.ldmd[77312290]: child 77304103 exited with 
> status 1
> Aug  9 13:28:21 6Q:io ldm: LDM Stop (login name, tty, login time, 
> [from]): ldm        ttyq37       Aug  8 17:44  (132.208.133.
> 37)
> Aug 09 13:28:21 5Q:io rpc.ldmd[77312290]: Exiting
> Aug 09 13:28:21 5Q:io rpc.ldmd[77312290]: Terminating process group
> Aug 09 13:28:21 5Q:io striker[77959463]: Exiting
> Aug 09 13:28:21 5Q:io dns2[78016494]: Exiting
> Aug 09 13:28:21 5Q:io 128.174.80.47[71288800]: Exiting
> Aug 09 13:28:21 5Q:io pqact[77024855]: Exiting
> Aug 09 13:28:21 5Q:io flood[77628265]: Exiting
> Aug 09 13:28:21 5Q:io pqbinstats[77339594]: Exiting
> 
> 
> Christian Pagé
> UQAM
> 
> 

===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================