Re: [Bacula-users] Using bconsole from command line
On Wed, 17 Oct 2007, Pascal Charest wrote: I am currently using bconsole for all my operations and would like to know if there is anyway I could script command from outside the bacula console. Below my .sig is the script I use to print hard copies of the jobs found on tapes that go into secure storage. My thinking is that if everything but the tapes gets destroyed, I'd at least have an inkling of which tapes to use first. -- Paul Heinlein [EMAIL PROTECTED] http://www.madboa.com/ #!/bin/bash # # Script for printing list of jobs on any given Bacula volume. # # == VOL=$1 TYPE= PFX=/usr/local/bacula case $VOL in '') cat __eom__ /dev/stderr This script sends to the printer a list of jobs stored on any given bacula tape volume. The printed output should be stored in the tape case when a volume is moved to the safe or off-site. Usage: $(basename $0) volume-name Example: $(basename $0) 06L3 __eom__ exit 1 ;; *) # pools are named daily-tape, weekly-tape, and monthly-tape, so # the $TYPE of job is discerned by stripping '-tape' from the # end of the Pool.Name. SQL= select Pool.Name from Media, Pool where Media.VolumeName = '$VOL' and Media.PoolId = Pool.PoolId; TYPE=$(echo $SQL | mysql bacula | tail -n1 | sed s/-tape$//) ;; esac test -n $TYPE || exit 1 ${PFX}/sbin/bconsole -c /etc/bacula/bconsole.conf \ __eof__ | a2ps --columns=1 --chars-per-line=132 @# @# * @# Listing $TYPE jobs on volume $VOL @# * @# @output /dev/null @# the initial output and menu just clutter things up query 14 @output $VOL quit __eof__ ### ### eof ### - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Reality check
I recently migrated from an LTO-1 drive (in a PowerVault 122T) to an LTO-3 (PowerVault 124T). Bacula (2.0.3) is locally configured to spool up to 50 GB to disk before writing to tape. Bacula reports that transfer rates have risen from about 22 Mbytes/second on the LTO-1 to about 43 Mbytes/second on the LTO-3. Does anyone have logs lying around that might confirm or deny whether those numbers are in the range of reasonable? -- Paul Heinlein [EMAIL PROTECTED] http://www.madboa.com/ - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Reality check
On Tue, 9 Oct 2007, John Drescher wrote: On 10/9/07, Paul Heinlein [EMAIL PROTECTED] wrote: I recently migrated from an LTO-1 drive (in a PowerVault 122T) to an LTO-3 (PowerVault 124T). Bacula (2.0.3) is locally configured to spool up to 50 GB to disk before writing to tape. Bacula reports that transfer rates have risen from about 22 Mbytes/second on the LTO-1 to about 43 Mbytes/second on the LTO-3. Does anyone have logs lying around that might confirm or deny whether those numbers are in the range of reasonable? Although these numbers are highly dependent on a lot of factors (compression, filesystem performance, database, network speed) which make comparing numbers not exactly perfect however your numbers look pretty good to me. I understand the caveats about equating one network's numbers with another's -- but thanks for letting me know mine are at least within a standard deviation of reasonable. :-) -- Paul Heinlein [EMAIL PROTECTED] http://www.madboa.com/ - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] RESOLVED: Re: UTF-8 file names support in Bacula ( OS X client)
On Fri, 30 Jun 2006, Frank Sweetser wrote: On Fri, Jun 30, 2006 at 05:03:37PM +0200, Kern Sibbald wrote: Egads, thanks for the details. I had forgotten about that aspect of Unicode since I rarely work with Windows (or Mac). Hopefully myself or someone else Frank? could summarized this for the manual. For the moment, the manual lacks all mention of Unicode/UTF-8, so I will add this to my todo so it does not get lost. I've thrown a note in the wiki about adding this in. I'll make sure to try and write something up when I get back in a few weeks. On a fairly recent Linux workstation, you can generate some UTF-8 filenames fairly easily: mkdir utf8-test cd utf8-test LANG=en_US.UTF-8 # get a list of UTF-8 locales on this system for loc in $(locale -a | grep utf8 | sort); do # generate filenames of days of the week # in various locales for n in $(seq 1 7); do touch $(env LANG=$loc date +%A -d 2006/7/$n) done done ls Test by backing up and restoring that directory with Bacula. -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnkkid=120709bid=263057dat=121642 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Re: [Bacula-devel] Fwd: Time for a Bacula book?
On Thu, 12 Jan 2006, Andrew Ford wrote: Maybe what is needed is a separate documentation mailing list to gather together the interested parties and then to draw up a strategy for the documentation as a sub-project. If part of the move includes migrating the documentation to something like DocBook XML (a good idea, imo), I'm willing to volunteer some time. The migration would likely encompass markup of the existing docs and assembling a decent toolchain for building various output formats: (x)html, man pages, PDF, ... -- Paul Heinlein Galois Connections, Inc. Systems Administrator 12725 SW Millikan Way, Ste. 290 [EMAIL PROTECTED] Beaverton, Oregon 97005 +1 503 626-6616 x140 www.galois.com --- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://ads.osdn.com/?ad_id=7637alloc_id=16865op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Low transfer speed to tape
On Fri, 28 Oct 2005, Gillow, Jeremy wrote: I'm seeing very low throughput when doing a full backup to DDS-4 tape from one Solaris machine to another over 100Mbit ethernet. When using rmt I typically get about 1.5MB/sec transfer rate, and when using bacula the throughput varies between 100KB/sec to 450KB/sec with an average of 350KB/sec. Average CPU usage for bacula-sd was 10% and the average for bacula-fd was 8%. For example, a system that normally takes 30 minutes to back up fully took 2 hours with bacula. I may be stating the obvious. If so, I apologize in advance... The Rate of backup reported by bacula is the number of bytes divided by the total time (in seconds) the backup job takes. The job may include any number of components: * stuff run before the job * identifying files to be archived * spooling the data to storage * spooling attributes to the directory * stuff run after the job The log snippet below doesn't include any run-before/after tasks so it's fairly simple. It takes about 10 1/2 minutes to run: * 3 minutes to spool files to disk, then tape * 7 minutes to spool attributes to database (The times are approximate within one minute.) 27-Oct 23:47 xxx-dir: Start Backup JobId 696, Job=yyy-backup.2005-10-27_23.30.07 27-Oct 23:47 xxx-sd: Spooling data ... 27-Oct 23:49 xxx-sd: Committing spooled data to Volume. Despooling 1,231,664,896 bytes ... 27-Oct 23:50 xxx-sd: Sending spooled attrs to the Director. Despooling 7,634,852 bytes ... 27-Oct 23:57 xxx-dir: Bacula 1.36.3 (22Apr05): 27-Oct-2005 23:57:56 JobId: 696 Job:yyy-backup.2005-10-27_23.30.07 Backup Level: Incremental, since=2005-10-27 00:31:05 Client: yyy-fd FileSet:yyy-fileset 2005-08-12 23:45:03 Pool: daily-tape Storage:PV122T Start time: 27-Oct-2005 23:47:14 End time: 27-Oct-2005 23:57:56 FD Files Written: 23,597 SD Files Written: 23,597 FD Bytes Written: 1,225,548,154 SD Bytes Written: 1,229,580,582 Rate: 1909.0 KB/s The Rate would be much higher if it included only the time needed to spool files to tape; it's the attribute spooling that lowers the average B/s. -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com --- This SF.Net email is sponsored by the JBoss Inc. Get Certified Today * Register for a JBoss Training Course Free Certification Exam for All Training Attendees Through End of 2005 Visit http://www.jboss.com/services/certification for more information ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] ClientRunBeforeJob problem
On Thu, 27 Oct 2005, Josh Fisher wrote: I have a problem with the following ClientRunBeforeJob script running on a Fedora Core 3 client machine using bacula-fd version 1.36.3. #!/bin/sh ldapbak=/etc/bacula/ldap_backup.ldif rm -f $ldapbak /etc/init.d/ldap stop /dev/null slapcat -l $ldapbak ret=$? /etc/init.d/ldap start if [ $ret != 0 ]; then exit 1 fi exit 0 On the client machine, bacula-fd runs as user root and group bacula. The slapcat command fails when the job is started from bconsole, returns slap_startup failed message, and cancels the job due to the script returning non-zero. However, I can manually run the script from a shell on the client machine without issue. Any idea what I'm doing wrong? Instead of invoking the init script directly, try using the /sbin/service wrapper script. It'll setup your environment correctly, which might not be happening in the bacula app space. Instead of /etc/init.d/ldap {start,stop} do /sbin/service ldap {start,stop} -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com --- This SF.Net email is sponsored by the JBoss Inc. Get Certified Today * Register for a JBoss Training Course Free Certification Exam for All Training Attendees Through End of 2005 Visit http://www.jboss.com/services/certification for more information ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] *Extreme* low MySQL performance when prunnig or purging jobs/volumes
On Wed, 26 Oct 2005, Russell Howe wrote: Check to see if Postgres is issuing an fsync() after every operation or not, and then decide if you need it to. If you can disable that, things should get a *LOT* faster. I posted a link to some Postgresql performance tuning documents a while back - they're linked to from the postgres website too. At the risk of being somewhat off-topic, I'll add that the same is true of SQLite. In SQLite 2.8, setting the default_synchronous flag to OFF will speed things up markedly: echo 'PRAGMA default_synchronous=0;' | sqlite bacula.db The danger is that a system crash can easily result in a wholly corrupted database. In fact, the SQLite developers have removed this functionality from version 3, calling it dangerous. Make sure you read the documentation to understand the dangers: http://www.sqlite.org/pragma.html The upside is that attribute spooling that used to take a couple days is now done in a few hours. My weekly full backups used to take from Friday night until Monday night; now they start Saturday evening and are finished about the time I get my Sunday-morning coffee. -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com --- This SF.Net email is sponsored by the JBoss Inc. Get Certified Today * Register for a JBoss Training Course Free Certification Exam for All Training Attendees Through End of 2005 Visit http://www.jboss.com/services/certification for more information ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bitrate problems using Bacula
On Mon, 29 Aug 2005, Romain wrote: Now I have bacula backuping 30 servers and it seems to work fine. I backup 1,5GB a day and the total file retention is near 52 Gb. The problem is that all backups seem to run very slow. The average bitrate I have is between 600 and 700 kb/s. But all my servers are plugged on 100Mb switches and the backup server is composed with 3 RAID5 146 Gb 10kt SCSI disks (note : all backups are made direct to disk). My experience is that the bottleneck is not archiving the files -- to tape or file or whatever -- but inserting the file attributes into the storage database. Worst case on our network is our cvs/subversion server that hosts 2.2 million files taking up only 90 GB of space. It takes only 4 hours to write the data to an LTO-1 tape drive. It then takes 26 hours to insert 715 MB of file attributes into the sqlite database. The archiving operation is quite fast, ca. 6.4 MB/s, but the overall rate is only in the range of 820 KB/s. Best case for us is our mail server, which still uses the old mailbox storage, with many messages per file, so we back up lots of data in relatively few files. With far less attribute spooling, these backups run 3.5 to 4 times faster (overall) than the worst case. My suggestion would be for you to identify the local bottleneck a bit more thoroughly. If the database operations are slowing you down, you might consider moving to a more efficient RDBMS or beefing up your server (faster disks, more RAM, ...). -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com --- SF.Net email is Sponsored by the Better Software Conference EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile Plan-Driven Development * Managing Projects Teams * Testing QA Security * Process Improvement Measurement * http://www.sqe.com/bsce5sf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Excludes not working?
On Thu, 25 Aug 2005, j2 wrote: With the below fileset, I thought all gzipped logs would be excluded from my backup, but as shown at the bottom of the post, they are not? [] regexfile = /var/log/.*\.[0-9]\.gz My guess (and it's only a guess) is that '.*' won't match a '/'; you'll need to specify the depth, e.g., regexfile = /var/log/.*\.[0-9]\.gz regexfile = /var/log/.*/.*\.[0-9]\.gz -- Paul Heinlein [EMAIL PROTECTED] www.madboa.com --- SF.Net email is Sponsored by the Better Software Conference EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile Plan-Driven Development * Managing Projects Teams * Testing QA Security * Process Improvement Measurement * http://www.sqe.com/bsce5sf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
RE: FW: [Bacula-users] bconsole spawns thousands of copies of itself
On Wed, 10 Aug 2005, Ron Wheeler wrote: The fact that it crashes on -? and the eror in building conio is leaningthis way. I do not know why the build is successful if the compile of conio fails. That looks like a bug in the make. I'm a recent subscriber to bacula-users, so forgive me if using strace has already been suggested... strace -f -o /tmp/bconsole.trace ./bconsole -? The output in /tmp/bconsole.trace should provide some clue as to what system call is tripping up. Also, it might be worthwhile running ldd against your binary to make sure you're linked against the right libraries, e.g., # ldd /usr/sbin/bconsole libtermcap.so.2 = /lib/libtermcap.so.2 (0x0099) libpthread.so.0 = /lib/tls/libpthread.so.0 (0x002ca000) libstdc++.so.5 = /usr/lib/libstdc++.so.5 (0x005a2000) libm.so.6 = /lib/tls/libm.so.6 (0x00bfa000) libgcc_s.so.1 = /lib/libgcc_s.so.1 (0x0057f000) libc.so.6 = /lib/tls/libc.so.6 (0x00dfa000) /lib/ld-linux.so.2 = /lib/ld-linux.so.2 (0x004af000) -- Paul Heinlein [EMAIL PROTECTED] --- SF.Net email is Sponsored by the Better Software Conference EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile Plan-Driven Development * Managing Projects Teams * Testing QA Security * Process Improvement Measurement * http://www.sqe.com/bsce5sf ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users