Conference turris::digital_unix

8807.0. "v3.2d2 cron dumps core ....." by BACHUS::DIDDEN () Wed Feb 12 1997 13:40

    We are having a problem with cron dumping core.
    
    I ran dbx and found the following trace.....
    
    
    tscosf_#file core_mob
    core_mob:       core dump, generated from 'cron'
    tscosf_#dbx /usr/sbin/cron core_mob
    dbx version 3.11.8
    Type 'help' for help.
    Core file created by program "cron"
    
    warning: /usr/sbin/cron has no symbol table -- very little is supported
    without
    it
    
    
    signal Segmentation fault at >*[NLstrdlen, 0x3ff800c7810]       bis    
    r1, r2,
    r1
    (dbx) t
    >  0 NLstrdlen(0x140001698, 0x1400018a0, 0x11ffff330, 0x100000018,
    0x3ff800cd490
    ) [0x3ff800c7810]
       1 _doprnt(0x140000251, 0x11ffffb30, 0x28, 0x7e04, 0x3ff800ece70)
    [0x3ff800c4f
    38]
       2 sprintf(0x140001af8, 0x140000230, 0x14000ab40, 0x140000258,
    0x53300d6c8) [0
    x3ff800c7914]
    (dbx) quit
    
     I checked the patchfile but found nothing ,looked around in
     COMET and found the following entry but it is without a solution,
     can anyone give a advise on this.
    
    thanks,
    
    Geert\\\
    
    
    ============================ COMET ==================================
    
    S:SRC961007002143
    B:16600
    I:US SURGICAL CORP
    U:NORWALK
    R:CT
    C:ERNEST BODURTHA
    A:745004
    T:(203)8345225      
    P:dunix - cron intermittently stops for no reason, nothing logged
    N:0001
    \+-----------------------  Beginning of Screener's Text
    -----------------------+
    
    \<s1> unix 3.2d cron stops from time to time. gt74408<cb 8-OCT
    13:13mdt> cb2
    \elv s1 jd 24773
    \+--------------------------- End of Screener's Text
    --------------------------+
    
    Will check for patches.
    
    Gene 32283
    N:0002
    Sent following mail to customer:
    
    Subject:
                C961007-2143 cron dies
           Date:
                Thu, 10 Oct 1996 08:58:57 -0400
          From:
                Gene Zurik <zurik@decatl.alf.dec.com>
    Organization:
                CSC
             To:
                ebodurtha@ussurg.com
    
    
    Ernest,
    
    I could not locate any patches for your problem but I did see
    where another customer had a similiar problem.
    
    Please mail me the following files.
    
    The output of crontab -l
    
    var/adm/cron/log     around the time of failure
    
    /var/adm/syslog.dated/date/daemon.log
    
    for date- use date of failure or date where daemon.log has data
    
    uerf -r 300 -R     one startup entry so I can see whats' on
                       your system
    
    Any other output from uerf near time of error
    
    or anything else you think may be of help.
    
    Thanks
    
    
    N:0003
    First mail did not go. Customer has new mail addr.
    
    ernest.bodurtha@qmgate.ussurg.com
    
    Resent mail.
    
    Gene 32283
    
    N:0004
    \+-----------------------  Beginning of Screener's Text
    -----------------------+
    
    \<s1>re C961007-2143 dunix - cron intermittently stops for no reason,
    nothing
    \logged. mr31524
    \+--------------------------- End of Screener's Text
    --------------------------+
    
    Called.  Jene and left voice.  I am working the same call.
    
    See call seq# C961028-1364.
    
    
    31050 Pat
    N:0005
    See last w/u for additional information.
    
    I spoke with the customer and he says none of the files mentioned on
    gene's
    mail contain any information. All he found was a core file. The OSAPPL
    team
    took a look at this core file and found the following:
    
    decatl.alf.dec.com> dbx /usr/sbin/cron ./core.ussurgical
    dbx version 3.11.8
    Type 'help' for help.
    Core file created by program "cron"
    
    warning: /usr/sbin/cron has no symbol table -- very little is supported
    without
    it
    
    
    signal Segmentation fault at >*[NLstrdlen, 0x3ff800c7800]       ldq_u  
    r1, 0(r1
    
    6)
    (dbx) where
    >  0 NLstrdlen(0x0, 0x100000000, 0x11ffff3f0, 0x100000018,
    0x3ff800cd480) [0x3ff
    
    800c7800]
       1 _doprnt(0x140000251, 0x11ffffbf0, 0x28, 0x0, 0x3ff800ece60)
    [0x3ff800c4f28]
    
       2 sprintf(0x140001af8, 0x140000230, 0x14000aa00, 0x140000258,
    0x100362d30) [0
    
    x3ff800c7904]
    
    
    
    No patches for cron either...
    From Comet I found some other UNIX 4.0 CDE core files whicgh crash in a
    suspiciously similar manner....
    
    Display Magnified Output
    
    S:SRC960712003580
    B:154097
    I:CORNELL UNIVERSITY - 901
    U:ITHACA
    R:NY
    C:DSN%LNS592:: SEB
    A:70076
    T:(000)0000000
    P:(pr1) segmentation fault in dxpause (security)
    N:0001
    N:0001
    dbx shows the following.
    ====================================================
    lnsb03.lns.cornell.edu> dbx /usr/bin/X11/dxpause core
    dbx version 3.11.10
    Type 'help' for help.
    Core file created by program "dxpause"
    
    warning: /usr/bin/X11/dxpause has no symbol table -- very little is
    supported without it
    
    
    signal Segmentation fault at >*[NLstrdlen, 0x3ff800d1880]       ldq_u  
    r1,
    0(r16)
    (dbx)
    
    
    
    Problem turned out to be with /usr/shlib/libXimp.so. If the library was
    renamed then dxpause went on without complaining and would not crash. I
    wonder if this is some sort of Internationalization bug?
    
    Called customer an this cron job doesnt do anything with libXimp.so,
    also
    the cron image s=doesn seem to be linked with this, so I think this may
    be
    a read herring.
    
    I will see what can be done about obtaining a cron with symbols to get
    a
    better traceback next time.
    
    Customer wanted to know if there is a way to get the daemon logs to
    repot
    more detail on cron's activity. He doesnt even see the jobs being
    started.
    
    Sent email to Gene.
    
    31050 pat
    N:0006
    Have not heard from customer. Assuming problem is fixed.
    
    Gene 32283