Hi again,
I've just finished to make another test:
the sleep process in mysqladmin processlist is the status query command
that i run when I run status director command from bconsole and that is
blocked when it starts showing me the running jobs (the future schedule
jobs are shown correctly, but it will never runs if I don't restart
director).
Here's the point where it's blocked
Scheduled Jobs:
Level Type Pri Scheduled Name Volume
===================================================================================
Full Backup 9 16-Jul-07 19:00 job-catalog
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-oracolo
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-samba3
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-etrust
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-caronte
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-proxy
Volume-0004
Differential Backup 10 16-Jul-07 20:00 job-ads
Volume-0004
Full Backup 10 16-Jul-07 22:00 job-ced
Volume-0004
Full Backup 10 16-Jul-07 23:00 job-nt4
Volume-0004
Differential Backup 10 16-Jul-07 23:00 job-anagrafe
Volume-0004
Differential Backup 10 16-Jul-07 23:00 job-neo
Volume-0004
Differential Backup 10 16-Jul-07 23:00 job-llpp
Volume-0004
Differential Backup 10 16-Jul-07 23:00 job-mensa
Volume-0004
Differential Backup 10 16-Jul-07 23:00 job-vigili
Volume-0004
====
Running Jobs:
--- the cursors blocked here ---
when i press CTRL-c the mysqladmin processlist mantain that process in
sleep mode (same ID).
only restart the time at 1.
Thanks
Bye
Alfredo Marchini wrote:
Hi,
bacula-dir is blocked again:
I've just made this tests:
- time command from bconsole: works
- mysqladmin processlist:
there are 2 processes:
+-----+--------+-----------+--------+---------+------+-------+------------------+
| Id | User | Host | db | Command | Time | State |
Info |
+-----+--------+-----------+--------+---------+------+-------+------------------+
| 275 | bacula | localhost | bacula | Sleep | 793 |
| |
| 276 | root | localhost | | Query | 0 | | show
processlist |
+-----+--------+-----------+--------+---------+------+-------+------------------+
i've also setted yesterday debug at 400, strace on and launched strace
program.
I attach my log files, but I cannot attach strace log files, because
they are too large (7.5Mb).
I've launched strace in background with these options:
strace -ff -T -o /root/strace-dir/bacula-dir.log -p 5540
Also I attach bacula-dir log.
Reading director.log, I think that there is a bug.
So strace can be very useful, but it's also very large.
I can publish strace log files, if needed, on my private web-server.
If bacula developers or somelse needs to make tests on director or
other info,
I can keep bacula-dir blocked without restarting for some days again.
Thanks
Alfredo
Arno Lehmann wrote:
Hi,
05.07.2007 21:10,, Alfredo Marchini wrote::
...
about 2 weeks and the DIR will block again.
I'll wait for it.
Ok, that sounds reasonable.
strace has many options, which I need to specify to view all the
information useful to find the problem?
Then I recommend you start playing with strace now :-)
The simples form is 'strace -p PID' which attaches to the process
with the given PID and reports lots of (library/kernel) calls.
In case a program is stuck, you only see the last call that the
programm is waiting for.
In many cases, that will tell you something already, like when a
programm is reading from a file without data.
In Bacula, you will often see repeating messages related to timers,
non-blocking socket operations, and so on, so you need the strace
before things start to repeat without changes.
The real problem is that I don't know which cause can create the lock.
I'm desperate because I don't have any idea, and all the
configuration files seems to be well configured,
or not configured to cause this problem.
Looks like it, yes.
Turning on debugging and sending the output to a trace file will be
helpful. Also, try it in the next weeks to see how normal operations
run, and then you'll probably notice something that either simply
gets stuck or repeats and repeats or even gets an error with
description.
Arno
------------------------------------------------------------------------
14-Jul 09:18 oracolo-director: Fatal Error at bnet_server.c:172 because:
Error in select: Unknown error 514
14-Jul 09:18 oracolo-director: ABORTING due to ERROR in smartall.c:193
qp->qnext->qprev != qp called from dlist.c:341
14-Jul 09:18 oracolo-director: Fatal Error because: Bacula interrupted by
signal 11: Segmentation violation
14-Jul 09:18 oracolo-director: Fatal Error at bnet_server.c:172 because:
Error in select: Unknown error 514
14-Jul 09:18 oracolo-director: ABORTING due to ERROR in smartall.c:193
qp->qnext->qprev != qp called from dlist.c:341
14-Jul 09:18 oracolo-director: Fatal Error because: Bacula interrupted by
signal 11: Segmentation violation
------------------------------------------------------------------------
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
------------------------------------------------------------------------
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
--
Alfredo Marchini
Consulente IT
P.IVA: 05649240487
CF: MRCLRD81R07D612B
Via Imbriani, 66
50019 Sesto Fiorentino (FI)
Tel. +39 393 9566375
E-Mail: [EMAIL PROTECTED]
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
