On Saturday 15 September 2007 16:44, Dan Langille wrote: > FYI, SD is crashing on 2.2.4 for this chap. See other email for > backtrace.
Some wild speculation with little info, starting with the less probable: 1. There is some hardware or driver problem with his tape drive. The particular tape may also be bad or blank (it may need an eof written). 2. His build is bad. 3. A binary built on one OS/architecture was moved to another different OS/architecture. What is clear is that Bacula died due to a buffer overrun -- something I have never seen in the SD, so this is a very unusual case. Since there is no info on what system he is running, or his Bacula .configure, or the other crashes, I cannot make any more comments. Kern > > ------- Forwarded message follows ------- > Date sent: Sat, 15 Sep 2007 14:43:20 +0200 > From: Marc Schiffbauer <[EMAIL PROTECTED]> > To: bacula-users <[EMAIL PROTECTED]> > Subject: Re: [Bacula-users] bacula-sd always crashing on job > start > > [ Double-click this line for list subscription options ] > > * Dan Langille schrieb am 15.09.07 um 05:04 Uhr: > > On 15 Sep 2007 at 4:59, Marc Schiffbauer wrote: > > Hi Dan, > > > > Hi, > > > > > > I have a problem with one of my SD's. > > > > > > Since upgrade from 2.0.3 to 2.2.3 it is always segfaulting when a > > > job starts to use it. > > > > I suggest upgrading to 2.2.4 released today, and check the release > > notes (see announce mailing list) for possible bug fixes related to > > this issue. > > I am using 2.2.4 now and it is still crashing. I found out that it > crashes even if I try to relabel a tape. > > I recompiled bacula and use an unstripped version of bacula-sd now. > Using the btrcaeback script I got the backtrace in the mail > attached. > > This is the console output: > [EMAIL PROTECTED]:~# bacula-sd -f -d50 -c /etc/bacula/bacula-sd.conf > lisa-sd: stored.c:480 SD init done /dev/nst0 > lisa-sd: stored.c:493 calling first_open_device "DLT8000" (/dev/nst0) > lisa-sd: cram-md5.c:71 send: auth cram-md5 > <[EMAIL PROTECTED]> ssl=0 > lisa-sd: cram-md5.c:71 send: auth cram-md5 <[EMAIL PROTECTED] > sd> ssl=0 > lisa-sd: dircmd.c:574 Found device DLT8000 lisa-sd: cram-md5.c:71 > send: auth cram-md5 > <[EMAIL PROTECTED]> ssl=0 lisa-sd: dircmd.c:574 Found > device DLT8000 > lisa-sd: cram-md5.c:71 send: auth cram-md5 > <[EMAIL PROTECTED]> ssl=0 > lisa-sd: cram-md5.c:71 send: auth cram-md5 > <[EMAIL PROTECTED]> ssl=0 > lisa-sd: dircmd.c:574 Found device DLT8000 > > Volume Label: > Id : Bacula 1.0 immortal > VerNo : 11 > VolName : Tape_13_2 > PrevVolName : > VolFile : 1 > LabelType : PRE_LABEL > LabelSize : 0 > PoolName : Default > MediaType : DLT-IV > PoolType : Backup > HostName : lisa > Date label written: 15-Sep-2007 14:27 > lisa-sd: reserve.c:313 jid=0 reserve_volume Tape_13_2 > lisa-sd: reserve.c:238 jid=0 new Vol=Tape_13_2 at 80d9910 > dev="DLT8000" (/dev/nst0) > lisa-sd: reserve.c:181 jid=0 List from end new volume: Tape_13_2 at > 80d9910 on device "DLT8000" (/dev/nst0) > 15-Sep 14:28 lisa-sd: ABORTING due to ERROR in smartall.c:215 > Buffer overrun called from mem_pool.c:213 > Kaboom! bacula-sd, lisa-sd got signal 11 - Segmentation violation. > Attempting traceback. > Kaboom! exepath=/root > Calling: /root/btraceback /root/bacula-sd 1493 > > ^^ this last line is wrong, and I called it manually as > /usr/bin/btraceback /usr/local/sbin/bacula-sd 1493 > > This is the output on bconsole: > > Connecting to Storage daemon lisa-sd at lisa:9103 ... > Sending relabel command from "Tape_13" to "Tape_13_2" ... > block.c:990 Read error on fd=8 at file:blk 0:0 on device "DLT8000" > (/dev/nst0). ERR=Cannot allocate memory. > 3000 OK label. VolBytes=102400 DVD=0 Volume="Tape_13_2" > Device="DLT8000" (/dev/nst0) > > > Is this useful? > > > If you are getting .core files, try supplying the output of this: > > > > # gdb bacula-sd bacula-sd.core > > backtrace > > > > FWIW, we've been working on a core of bacula-sd the past couple of > > days. > > Thanks for helping > -Marc > -- > begin LOVE-LETTER-FOR-YOU.txt.vbs > I am a signature virus. Distribute me until the bitter > end > > ---------------------------------------------------------------------- > --- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2005. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Bacula-users mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/bacula-users > > > ------- End of forwarded message ------- ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Bacula-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/bacula-devel
