Re: [Bacula-users] Second concurrent job fails if it requires another pool - bug?
On Thursday 10 August 2006 06:41, Ian Levesque wrote: Hi all, Here was the situation this evening: I'm backing up to a 3-drive LTO2 autochanger. There were six jobs running, filling the max clients per drive (set at 2 concurrent jobs each). All were full backups using a Fulls pool. Then, one of the full backups completes, opening the door for an incremental that was waiting for about 24 hours to run. When it finally gets its chance, it throws a fatal error because the volume mounted is in the Fulls pool, and it wants to use the Incrementals pool. What I don't understand is 1) why it doesn't wait for the other job to complete, rather than dying and 2) why the job started to begin with if the director knew that the wrong tape was mounted for the job. Is this a bug? Expected behavior? Hello, If you are still having problems with this, it probably is worth looking at. Please file a bug report, showing the Job output of failed jobs, and attach your Dir and your SD conf files to the bug report. Also, if you can, run the jobs with -d400 on the command line, capture the debug output, and add it as an attachment to the bug report. This kind of bug (if it is one) is not easy to solve, so don't expect any immediate resolution. Regards, Kern Full job, just completing: 09-Aug 20:59 sbgrid-sd: New volume 25L2 mounted on device LTO2B (/dev/nst1) at 09-Aug-2006 20:59. 10-Aug 00:19 sbgrid-dir: Bacula 1.38.11 (28Jun06): 10-Aug-2006 00:19:45 ... Incremental job, just starting: 10-Aug 00:19 sbgrid-dir: Start Backup JobId 308, Job=NMRL1.2006-08-09_06.00.00 10-Aug 00:14 nmrl1-fd: DIR and FD clocks differ by -350 seconds, FD automatically adjusting. 10-Aug 00:19 sbgrid-sd: NMRL1.2006-08-09_06.00.00 Fatal error: acquire.c:263 Wanted Volume 01L2, but device LTO2B (/dev/ nst1) is busy writing on 25L2 . Thanks for any ideas, and let me know if you need more logs, config details, etc. Ian - 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 - 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] Delete the oldest volume automatically
If I want only two volumes in a pool and each volume is used once. (one volume per day also) I put: Pool { Name = Automatico Pool Type = Backup Maximum Volumes = 2 Purge Oldest Volume = yes Maximum Volume Jobs = 1 LabelFormat = Copia_Postgres-${Day:p/2/0/r}-${Month:p/2/0/r}-${Year} } But when I third job runs, It will recycle the oldest volume, then the label will not change. I want that label with the correct date. Not a overwritted volume. How can I do that? Not an answer, but an additional warning: Be careful that your data will definety fit into one volume, and that both the volumes are always available. Just in case, if something goes badly wrong with your system and it generates a lot of files, that don't any more fit into your one volume, the system will (AFAIK) immediatedly purge the older (yesterday's) volume to be able to finish the backup of all the files. Then you'll only have this one backup, which is from a system where are a lot of improperly generated files, and propably something else is also wrong there. And the worst case scenario is, if your data doesn't even then fit to the two volumes available -then it obviously purges the first volume (where the job was started to), and also the only backup you'll have will be sure to be invalid. Purge Oldest Volume = yes may be very dangerous. Since it is possible, I suggest using volume retention period of more than 24 but less than 48 hours, and pruning regular auto-recycling instead of auto-purge. More off-topic: I'm currently fighting with a slightly similar system, but using two (even/odd) monthly pools and trying to force a recycle in the beginning of every month. The problem there is that months are not of the same length, unlike the days are... -- TiN - 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
[Bacula-users] Can't get storage daemon configured
I'm trying to start up the storage daemon and I'm getting an error: [EMAIL PROTECTED]:/var/lib/bacula# /etc/init.d/bacula-sd start Starting Bacula Storage daemon: 26-Aug 07:34 bacula-sd: ERROR TERMINATION at parse_conf.c:821 Config error: Keyword DeviceType not permitted in this resource. Perhaps you left the trailing brace off of the previous resource. : line 47, col 13 of file /etc/bacula/bacula-sd.conf DeviceType = DVD Here is my bacula-sd.conf: Storage { # definition of myself Name = r2d2-sd SDPort = 9103 # Director's port WorkingDirectory = /var/lib/bacula Pid Directory = /var/run/bacula Maximum Concurrent Jobs = 20 } Director { Name = r2d2-dir Password = xxx } Director { Name = r2d2-dir Password = xxx } Device { Name = DVDStorage DeviceType = DVD Media Type = DVD Archive Device = /dev/hdc LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = yes AlwaysOpen = no Spool Directory = /bacula-spool Requires Mount = Yes Mount Point = /bacula-mount Mount Command = /bin/mount -t iso9660 -o ro %a %m Unmount Command = /bin/umount %m Write Part Command = /usr/local/bin/dvd-handler %a write %e %v Free Space Command = /usr/local/bin/dvd-handler %a free } - 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] backup extrem slow after upgrade
Just to add more probably relevant info: The backup medium is an external USB 2.0 disk. It has not changed in the time between the high performance with version 1.38.5 and the low performance with the current version 1.38.11. The DB is sqlite which served well form months before. I had to create it new because the new version did not start because of an DB version conflict. /m - 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] backup extrem slow after upgrade
On 8/26/06, Marco [EMAIL PROTECTED] wrote: Just to add more probably relevant info: The backup medium is an external USB 2.0 disk. It has not changed in the time between the high performance with version 1.38.5 and the low performance with the current version 1.38.11. The DB is sqlite which served well form months before. I had to create it new because the new version did not start because of an DB version conflict. Are you sure that the new database is indexed properly? Does it take an unreasonably long time to generate the file list for a restore? If so this is a sign that the database needs indexed. John - 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] how to take backup of open file OF WIN 2000, XP by bacula of LINUX
Hello, On 8/24/2006 12:48 PM, parveen wrote: hi all i have installed bacula on linux. i am not able to take backup of open file (microsoft outlook express, VB files etc) plz give me some solution. The keyword to look for is VSS. Set up your windows jobs to use VSS and everything should work. I don't know if VSS is available for Win2000, though. Arno warm regrads parveen kumar Tetra Information Services Pvt. Ltd 136, Lower ground, Sant Nagar East of Kailash New Delhi -110065 Phone : 55604033,55604034,55604035 Fax : 26225293 www.linux4e.com,www.tetrain.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 -- IT-Service Lehmann[EMAIL PROTECTED] Arno Lehmann http://www.its-lehmann.de - 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] backup extrem slow after upgrade
John Drescher wrote: Are you sure that the new database is indexed properly? Does it take an unreasonably long time to generate the file list for a restore? If so this is a sign that the database needs indexed. The restore time is quite ok. It's the backup that takes extremely long. /m - 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
[Bacula-users] IPv6 host resolution
Hi list, I'm trying to set up Bacula to use my IPv6 network to get around the problem that for some of my backup clients over IPv4, the storage daemon is behind NAT. On all of the hosts involved, I've set up the daemons to only listen on IPv6 addresses with the exception of the director which could still service requests to/from local IPv4 clients, and all of the config for connections to any of the other daemons use the hostname for the Address = parameter as I'm also using TLS with the commonName of the certificate matching the hostname of the server. I have DNS records set up correctly such that ping hostname and ping6 hostname both work as expected from all of the hosts involved. The problem I have is, Bacula seems to always resolve and use the IPv4 address of the host before the IPv6 address. On the host running the director daemon, I am also running one of the DNS servers authoritative for the domain of these hosts, so I can see the DNS queries being made, and I only see queries for 'A' records and no '' records. So for example, if I run bconsole and try status client and pick one of the clients only accessible over IPv6, it tries and fails to connect to the remote file daemon over IPv4. I wrote a small test program that uses the getaddrinfo(3) library call to return all of the addresses for a given hostname and when run on the host running the director daemon with the hostname of one of the remote hosts, it does what I expect and returns the IPv6 address first, followed by the IPv4 address. This is also true when run on any of the hosts involved. I can't find any reference to a standard that requires that IPv6 addresses should be returned before IPv4 ones, but every bit of software that is both IPv6 and IPv4 capable always seems to use the IPv6 address first. Is there a way to force Bacula to prefer IPv6 addresses or would this be considered a bug in the DNS resolving code? I had a look in the source (src/lib/bnet.c, right?) but it looks a bit more complicated than just calling one the C library calls and relying on the order from that, I'm not familiar enough with it (yet) to understand what's going on. At first glance it looks like in some cases the code fetches all of the IPv4 addresses followed by the IPv6 address, or only tries IPv6 resolution if IPv4 fails. Any ideas? Matt - 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] is mysql 5 supported
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 building and re-installing bacula-dir from a deb.src resolved the problem. Ivánszky Gábor wrote: Hi All, I have the same error messages on Debian etch with mysql 5.0.24-1, and bacula-director-mysql 1.38.11-1. Restarting doesn't seem to solve the problem. I switched on mysql logging, and only the connection made by bacula-dir can be seen, and not even a transaction or a disconnect :( I also made a tcpdump(attached). Seems that authentication works fine, but the first command from client is made with an invalid command code(81). Any ideas? regards, Gabor On Tuesday 08 August 2006 13:45, Silver Salonen wrote: Hi. I moved my Bacula to another machine (both FreeBSD-6.1 and Bacula-1.38.11). On an old machine there's MySQL 4.1.21, but on the new MySQL is 5.0.24. I moved bacula database over with mysqldump and got no errors during that. If I now try to start my new bacula-dir, it says: # /usr/local/etc/rc.d/bacula-dir start Starting bacula_dir. 08-Aug 12:46 bacula-dir: Fatal error: Version error for database bacula. Wanted 9, got 0 08-Aug 12:46 bacula-dir: Fatal error: Could not open Catalog MySQL, database bacula. 08-Aug 12:46 bacula-dir: Fatal error: Version error for database bacula. Wanted 9, got 0 08-Aug 12:46 bacula-dir ERROR TERMINATION Please correct configuration file: /usr/local/etc/bacula-dir.conf Is the problem in MySQL, database, data, Bacula, or bacula-dir's configuration? Any suggestions for fixing this? Thanks in advance, Silver It seems that the problem was in MySQL. Restarting mysql made it possible to start bacula-dir. Silver -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.2 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFE8OZdDiqVI6Xm21MRAliCAJsGnvCJSKYQev1fS4p39gE4ZSxBUgCgjv0G 2v92wjtTYu3UMxvmGKKhxHg= =hT0i -END PGP SIGNATURE- - 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