[Bacula-users] Maximum Network Buffer Size warning
This should probably work its way into the manual, but a warning for anyone who tries to move to significantly larger Maximum Network Buffer Size numbers: At least in bacula-1.38, Maximum Network Buffer Size _must_ be less than 51, or restores will crash the storage daemon. I was playing with Maximum Network Buffer Size and Maximum Block Size recently, and hit this problem. I naively assumed it would be beneficial to have 2 Maximum Block Size bytes of data buffered, and was trying to see if increasing the Maximum Block size helped my storage daemon throughput. Some testing with btape suggested that there was a performance benefit with larger block sizes (although since that isn't 'real' data it might not be correct). Anyway, I happily configured my bacula installation for 262144 maximum block size and 524288 maximum network buffer size (256K and 512K). Backups seemed to be fine, but when I tried to restore, the storage daemon kept crashing--I tracked it down to an ASSERT in src/stored/record.c (about line 494), which presumably exists to prevent a malformed block from chewing up too much memory. I've now changed my maximum network buffer size to less than 51, but now I'm uncertain what value for network buffer size leads to the best performance--it may be that it is best if it is a precise multiple of (Maximum Block Size - block overhead). At any rate, without a source change, restores won't work if it is more than 51--and, it may be that any performance gains from using a larger than default block size are not worth the trouble--at least now I have a work-around if I need to restore my most recent set of backups. -se --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] btape test works, btape fill doesn't
Hi,
I'm in the process of setting up bacula for the first time for our
network but am getting errors when I run btape fill. I'm running bacula
version 1.38.1 (installed from ports) on FreeBSD 5.4 with a Quantum SDLT
110/220 tape drive (which is inside a Overland LoaderXPress). If I run
the btape test command the output indicates everything worked
successfully, however when running btape fill using single tape mode
btape gives me the following error:
btape: btape.c:2329 End of tape 108:0. VolumeCapacity=102,275,630,752.
Write rate = 9689.8 KB/s
Done writing ...
Done filling tape at 108:0. Now beginning re-read of tape ...
ERROR! device at 1:1 count=1
ERROR! device at 1:1 count=1
Reposition from 1:1 to 106:10374
I waited about 30 minutes after receiving this last message but btape
didn't display anything else. Assuming it had hung I killed it with
+.
Output from btape fill, btape test and my bacula-sd.conf file are
included below.
Does anybody have any ideas about what might be wrong here? I'm trying
to get bacula into production in the next few days, so any help would be
greatly appreciated.
Thanks,
Joe
Here's the device definition from my bacula-sd.conf file:
Device {
Name = Overland
Description = "Overland on FreeBSD"
Media Type = SDLT-1
Archive Device = /dev/nsa0
AutomaticMount = yes; # when device opened, read it
AlwaysOpen = yes
Offline On Unmount = no
Hardware End of Medium = no #FreeBSD secific
BSF at EOM = yes #FreeBSD secific
Backward Space Record = no#FreeBSD secific
Fast Forward Space File = no #FreeBSD secific
TWO EOF = yes #FreeBSD secific
# Autochanger = yes
# Changer Device = /dev/pass0
# Changer Command = "/usr/local/sbin/mtx-changer %c %o %S %a %d"
}
The device has an autochanger but I want to get it working without an
autochanger first.
Here's some of the output from btape fill:
[EMAIL PROTECTED] btape /dev/nsa0
Tape block granularity is 1024 bytes.
btape: butil.c:266 Using device: "/dev/nsa0" for writing.
btape: btape.c:338 open device "Overland" (/dev/nsa0): OK
*fill
This command simulates Bacula writing to a tape.
It requires either one or two blank tapes, which it
will label and write.
If you have an autochanger configured, it will use
the tapes that are in slots 1 and 2, otherwise, you will
be prompted to insert the tapes when necessary.
It will print a status approximately
every 322 MB, and write an EOF every 3.2 GB. If you have
selected the simple test option, after writing the first tape
it will rewind it and re-read the last block written.
If you have selected the multiple tape test, when the first tape
fills, it will ask for a second, and after writing a few more
blocks, it will stop. Then it will begin re-reading the
two tapes.
This may take a long time -- hours! ...
Do you want to run the simplified test (s) with one tape
or the complete multiple tape (m) test: (s/m) s
Simple test (single tape) selected.
Wrote Volume label for volume "TestVolume1".
Wrote Start of Session label.
Begin writing Bacula records to tape ...
Wrote blk_block=5000, dev_blk_num=4999 VolBytes=322,495,472 rate=10078.0
KB/s
Wrote blk_block=1, dev_blk_num= VolBytes=645,055,424
rate=10239.0 KB/s
Wrote blk_block=15000, dev_blk_num=14999 VolBytes=967,615,368
rate=10293.8 KB/s
Flush block, write EOF
Wrote blk_block=2, dev_blk_num=5000 VolBytes=1,290,175,312
rate=10001.4 KB/s
Wrote blk_block=25000, dev_blk_num=1 VolBytes=1,612,735,264
rate=10079.6 KB/s
Wrote blk_block=3, dev_blk_num=15000 VolBytes=1,935,295,208
rate=9874.0 KB/s
Flush block, write EOF
...
Wrote blk_block=155, dev_blk_num=5000 VolBytes=99,993,519,128
rate=9695.9 KB/s
Wrote blk_block=1555000, dev_blk_num=1 VolBytes=100,316,079,072
rate=9697.1 KB/s
Wrote blk_block=156, dev_blk_num=15000 VolBytes=100,638,639,016
rate=9699.2 KB/s
Flush block, write EOF
Wrote blk_block=1565000, dev_blk_num=5000 VolBytes=100,961,198,968
rate=9692.9 KB/s
Wrote blk_block=157, dev_blk_num=1 VolBytes=101,283,758,912
rate=9695.0 KB/s
Wrote blk_block=1575000, dev_blk_num=15000 VolBytes=101,606,318,864
rate=9697.1 KB/s
Flush block, write EOF
Wrote blk_block=158, dev_blk_num=5000 VolBytes=101,928,878,808
rate=9695.5 KB/s
Wrote blk_block=1585000, dev_blk_num=1 VolBytes=102,251,438,752
rate=9697.6 KB/s
14-Dec 17:20 btape: End of Volume "TestVolume1" at 106:10375 on device
"Overland" (/dev/nsa0). Write of 64512 bytes got 0.
btape: btape.c:2295 Last block at: 106:10374 this_dev_block_num=10375
btape: btape.c:2329 End of tape 108:0. VolumeCapacity=102,275,630,752.
Write rate = 9689.8 KB/s
Done writing ...
Wrote state file last_block_num1=10374 last_block_num2=0
Done filling tape at 108:0. Now beginning re-read of tape ...
Reading the first 1 records.
ERROR! device at 1:1 count=1
ERROR! device at 1:1 count=1
Reposition from 1:1 to 106:10374
Here is the output from btape
Re: [Bacula-users] Backup to file OK, but restore fails
Hi! Thank you Kern! This debug flag gave me a lot of information and also a hint to solve the problem! I have two storage resources (with file as media type) specified in the directors configuration file. One for a local file storage and one for the storage daemon on the second server. When doing a restore the director suggested the first of the two (the local one) and even though I manually changed the restore job to use the other sd the job failed. When looking at the debug output it became clear that the file daemon tried to connect to the wrong storage daemon to get the files. I commented out the local file storage and restarted the daemons. I could now do a restore without any problems! Is this a know issue? Howcome the file daemon tries to access the local storage daemon even though I changed "storage" to another sd when specifying the restore? I didn't see this problem before my upgrade. Could it be a bug in the new release or could I have a problem in the config somewhere? Is anyone else running several storage daemons with bacula 1.38.2? Best regards, Jonas Mixter On 2005-12-14 23:15, Kern Sibbald wrote: Hello Jonas, Using an IP address is fine. Running with a -d400 on the SD and the FD would probably give you a much better idea what is going wrong ... On Wednesday 14 December 2005 23:12, Jonas Mixter wrote: On 2005-12-14 22:23, Kern Sibbald wrote: On Wednesday 14 December 2005 22:13, Jonas Mixter wrote: On 2005-12-14 17:55, Attila Fülöp wrote: Jonas Mixter wrote: On 2005-12-14 10:29, Jonas Mixter wrote: Hi! Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that comes with Debian sarge) to 1.38.2. I have one machine running the director (with a connected tape-station) and one machine with a lot of disks and a storage daemon (this should later on be placed elsewhere, hence the separation of the daemons). The director is named merry-dir and the storage daemon is named pippin-sd. Before the upgrade I could backup for example the catalog to a file storage on the machine with just the sd. I could also restore files from the backup. After the upgrade, I could backup just as before. The job exits OK and I could see that the files that hold the backup are growing. But I cannot restore... When running a restore I could mark the files I want and the jobs is visible underneath "Running Jobs" in bconsole. The job is "waiting for Client merry-fd to connect to Storage File" and after a while the job times out. Here's the output when trying to restore a file to merry-fd (running on the same server as the director): 14-Dec 09:03 merry-dir: Start Restore Job RestoreFiles.2005-12-14_09.03.22 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Failed to authenticate Storage daemon. 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: Socket error on Storage command: ERR=No data available 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 0 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 1 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** Note the job fails twice... Is that really correct? I could backup and restore without any problems to a local file on the server running the director (tested only once though). I could check the status of the storage daemon from the director, and also do backups so the passwords should be OK, right? I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but don't really know the meaning. In a previous post i saw that Kern asked about the /lib/tls-directory when presented with this error. I do have the /lib/tls, but start all my daemons with the st
Re: [Bacula-users] quantum px502
Aditya Ivaturi wrote: > > We are looking to purchase a Quantum PX502 tape library with LTO-3 > drives. Any idea if this will be compatible with Bacula? I know of no reason it shouldn't be. -- Phil Stracchino [EMAIL PROTECTED] Renaissance Man, Unix generalist, Perl hacker Mobile: 603-216-7037 Landline: 603-886-3518 --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] quantum px502
We are looking to purchase a Quantum PX502 tape library with LTO-3 drives. Any idea if this will be compatible with Bacula? --Turi
Re: [Bacula-users] Backup to file OK, but restore fails
Hello Jonas, Using an IP address is fine. Running with a -d400 on the SD and the FD would probably give you a much better idea what is going wrong ... On Wednesday 14 December 2005 23:12, Jonas Mixter wrote: > On 2005-12-14 22:23, Kern Sibbald wrote: > >On Wednesday 14 December 2005 22:13, Jonas Mixter wrote: > >>On 2005-12-14 17:55, Attila Fülöp wrote: > >>>Jonas Mixter wrote: > On 2005-12-14 10:29, Jonas Mixter wrote: > >Hi! > >Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that > >comes with Debian sarge) to 1.38.2. > >I have one machine running the director (with a connected > > tape-station) and one machine with a lot of disks and a storage > > daemon (this should later on be placed elsewhere, hence the > > separation of the daemons). The director is named merry-dir and the > > storage daemon is named pippin-sd. Before the upgrade I could backup > > for example the catalog to a file storage on the machine with just > > the sd. I could also restore files from > >the backup. > > > >After the upgrade, I could backup just as before. The job exits OK > >and I > >could see that the files that hold the backup are growing. But I > > cannot restore... > >When running a restore I could mark the files I want and the jobs is > >visible underneath "Running Jobs" in bconsole. The job is "waiting for > >Client merry-fd to connect to Storage File" and after a while the job > >times out. > > > >Here's the output when trying to restore a file to merry-fd (running > > on the same server as the director): > > > >14-Dec 09:03 merry-dir: Start Restore Job > >RestoreFiles.2005-12-14_09.03.22 > >14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >Authorization key rejected by Storage daemon. > >Please see > >http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors > >for help. > >14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >Failed to authenticate Storage daemon. > >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >Socket error on Storage command: ERR=No data available > >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula > >1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > >JobId: 512 > >Job:RestoreFiles.2005-12-14_09.03.22 > >Client: merry-fd > >Start time: 14-Dec-2005 09:03:25 > >End time: 14-Dec-2005 09:13:30 > >Files Expected: 1 > >Files Restored: 0 > >Bytes Restored: 0 > >Rate: 0.0 KB/s > >FD Errors: 0 > >FD termination status: SD termination status: Waiting on FD > >Termination:*** Restore Error *** > > > >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula > >1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > >JobId: 512 > >Job:RestoreFiles.2005-12-14_09.03.22 > >Client: merry-fd > >Start time: 14-Dec-2005 09:03:25 > >End time: 14-Dec-2005 09:13:30 > >Files Expected: 1 > >Files Restored: 0 > >Bytes Restored: 0 > >Rate: 0.0 KB/s > >FD Errors: 1 > >FD termination status: SD termination status: Waiting on FD > >Termination:*** Restore Error *** > > > > > >Note the job fails twice... Is that really correct? > >I could backup and restore without any problems to a local file on the > >server running the director (tested only once though). > >I could check the status of the storage daemon from the director, and > >also do backups so the passwords should be OK, right? > > > >I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job > >RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but > >don't > >really know the meaning. In a previous post i saw that Kern asked > > about the /lib/tls-directory when presented with this error. I do > > have the /lib/tls, but start all my daemons with the startscript that > > bacula provides. In that script the > > "LD_ASSUME_KERNEL=2.4.19"-variable is exported. (If running kernel > > 2.4! is the /lib/tls a problem with kernel 2.6 too?) > >I use kernel 2.4.27 on the server with the director, and 2.6.8 on > >the sd > >server. The /lib/tls is there on both the servers which are running > >Debian sarge. > >Should really "merry-sd" be involved? I've selected the storage on > >pippin-sd when running "restore" in bconsole. Could it be that the > >director/fd is trying to restore from the wrong sd? Shouldn't the job > >then fail with a "volume not found" error, ask me to mount a volume, > > or similiar? > > > >If
Re: [Bacula-users] Backup to file OK, but restore fails
On 2005-12-14 22:23, Kern Sibbald wrote: On Wednesday 14 December 2005 22:13, Jonas Mixter wrote: On 2005-12-14 17:55, Attila Fülöp wrote: Jonas Mixter wrote: On 2005-12-14 10:29, Jonas Mixter wrote: Hi! Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that comes with Debian sarge) to 1.38.2. I have one machine running the director (with a connected tape-station) and one machine with a lot of disks and a storage daemon (this should later on be placed elsewhere, hence the separation of the daemons). The director is named merry-dir and the storage daemon is named pippin-sd. Before the upgrade I could backup for example the catalog to a file storage on the machine with just the sd. I could also restore files from the backup. After the upgrade, I could backup just as before. The job exits OK and I could see that the files that hold the backup are growing. But I cannot restore... When running a restore I could mark the files I want and the jobs is visible underneath "Running Jobs" in bconsole. The job is "waiting for Client merry-fd to connect to Storage File" and after a while the job times out. Here's the output when trying to restore a file to merry-fd (running on the same server as the director): 14-Dec 09:03 merry-dir: Start Restore Job RestoreFiles.2005-12-14_09.03.22 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Failed to authenticate Storage daemon. 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: Socket error on Storage command: ERR=No data available 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 0 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 1 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** Note the job fails twice... Is that really correct? I could backup and restore without any problems to a local file on the server running the director (tested only once though). I could check the status of the storage daemon from the director, and also do backups so the passwords should be OK, right? I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but don't really know the meaning. In a previous post i saw that Kern asked about the /lib/tls-directory when presented with this error. I do have the /lib/tls, but start all my daemons with the startscript that bacula provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is exported. (If running kernel 2.4! is the /lib/tls a problem with kernel 2.6 too?) I use kernel 2.4.27 on the server with the director, and 2.6.8 on the sd server. The /lib/tls is there on both the servers which are running Debian sarge. Should really "merry-sd" be involved? I've selected the storage on pippin-sd when running "restore" in bconsole. Could it be that the director/fd is trying to restore from the wrong sd? Shouldn't the job then fail with a "volume not found" error, ask me to mount a volume, or similiar? If I try to cancel a job, the director from time to time complains it cannot the job even though it's listed under running jobs. *cancel Automatically selected Job: JobId=515 Job=RestoreFiles.2005-12-14_10.18.20 Confirm cancel (yes/no): yes 3902 Job RestoreFiles.2005-12-14_10.18.20 not found. The job is marked "has been canceled" in bconsole anyway. I don't remember if I got errors like this before the upgrade. What could be wrong in my setup? I've been banging my head to the wall for quite a few hours now and I'm out of ideas. Best regards, Jonas Mixter Hi! I've doing some more tests today. There seems to be no problems at all restoring from backups stored on tape. If I copy the entire backupfile from the server with running only the sd, to the server running the bacu
Re: [Bacula-users] Backup to file OK, but restore fails
On Wednesday 14 December 2005 22:13, Jonas Mixter wrote: > On 2005-12-14 17:55, Attila Fülöp wrote: > > Jonas Mixter wrote: > >> On 2005-12-14 10:29, Jonas Mixter wrote: > >>> Hi! > >>> Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that > >>> comes with Debian sarge) to 1.38.2. > >>> I have one machine running the director (with a connected tape-station) > >>> and one machine with a lot of disks and a storage daemon (this should > >>> later on be placed elsewhere, hence the separation of the daemons). The > >>> director is named merry-dir and the storage daemon is named pippin-sd. > >>> Before the upgrade I could backup for example the catalog to a file > >>> storage on the machine with just the sd. I could also restore files > >>> from > >>> the backup. > >>> > >>> After the upgrade, I could backup just as before. The job exits OK > >>> and I > >>> could see that the files that hold the backup are growing. But I cannot > >>> restore... > >>> When running a restore I could mark the files I want and the jobs is > >>> visible underneath "Running Jobs" in bconsole. The job is "waiting for > >>> Client merry-fd to connect to Storage File" and after a while the job > >>> times out. > >>> > >>> Here's the output when trying to restore a file to merry-fd (running on > >>> the same server as the director): > >>> > >>> 14-Dec 09:03 merry-dir: Start Restore Job > >>> RestoreFiles.2005-12-14_09.03.22 > >>> 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >>> Authorization key rejected by Storage daemon. > >>> Please see > >>> http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors > >>> for help. > >>> 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >>> Failed to authenticate Storage daemon. > >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: > >>> Socket error on Storage command: ERR=No data available > >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula > >>> 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > >>> JobId: 512 > >>> Job:RestoreFiles.2005-12-14_09.03.22 > >>> Client: merry-fd > >>> Start time: 14-Dec-2005 09:03:25 > >>> End time: 14-Dec-2005 09:13:30 > >>> Files Expected: 1 > >>> Files Restored: 0 > >>> Bytes Restored: 0 > >>> Rate: 0.0 KB/s > >>> FD Errors: 0 > >>> FD termination status: SD termination status: Waiting on FD > >>> Termination:*** Restore Error *** > >>> > >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula > >>> 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > >>> JobId: 512 > >>> Job:RestoreFiles.2005-12-14_09.03.22 > >>> Client: merry-fd > >>> Start time: 14-Dec-2005 09:03:25 > >>> End time: 14-Dec-2005 09:13:30 > >>> Files Expected: 1 > >>> Files Restored: 0 > >>> Bytes Restored: 0 > >>> Rate: 0.0 KB/s > >>> FD Errors: 1 > >>> FD termination status: SD termination status: Waiting on FD > >>> Termination:*** Restore Error *** > >>> > >>> > >>> Note the job fails twice... Is that really correct? > >>> I could backup and restore without any problems to a local file on the > >>> server running the director (tested only once though). > >>> I could check the status of the storage daemon from the director, and > >>> also do backups so the passwords should be OK, right? > >>> > >>> I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job > >>> RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but > >>> don't > >>> really know the meaning. In a previous post i saw that Kern asked about > >>> the /lib/tls-directory when presented with this error. I do have the > >>> /lib/tls, but start all my daemons with the startscript that bacula > >>> provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is > >>> exported. (If running kernel 2.4! is the /lib/tls a problem with kernel > >>> 2.6 too?) > >>> I use kernel 2.4.27 on the server with the director, and 2.6.8 on > >>> the sd > >>> server. The /lib/tls is there on both the servers which are running > >>> Debian sarge. > >>> Should really "merry-sd" be involved? I've selected the storage on > >>> pippin-sd when running "restore" in bconsole. Could it be that the > >>> director/fd is trying to restore from the wrong sd? Shouldn't the job > >>> then fail with a "volume not found" error, ask me to mount a volume, or > >>> similiar? > >>> > >>> If I try to cancel a job, the director from time to time complains it > >>> cannot the job even though it's listed under running jobs. > >>> *cancel > >>> Automatically selected Job: JobId=515 > >>> Job=RestoreFiles.2005-12-14_10.18.20 > >>> Confirm cancel (yes/no): yes > >>> 3902 Job RestoreFiles.2005-12-14_10.18.20 not found. > >>> > >>> The job is marked "has been canceled"
Re: [Bacula-users] clientrunbeforejob not working
On Wednesday 14 December 2005 21:04, Computer King wrote:
> I am tring to run the following 2 commands with bacula but them do not
> seem to work when i run them on the local machine it works fine however
> these commands must be run as root how can i get this to work with zimbra.
>
> ClientRunBeforeJob = /opt/zimbra/bin/zimbra stop
> ClientRunAfterJob = /opt/zimbra/bin/zimbra start
I suspect that you are missing quotes.
ClientRunBeforeJob = "/opt/zimbra/bin/zimbra stop"
ClientRunAfterJob = "/opt/zimbra/bin/zimbra start"
--
Best regards,
Kern
(">
/\
V_V
---
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=7637&alloc_id=16865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Error "waiting to reserve a device" after upgrade to 1.38.3
On Wednesday 14 December 2005 04:22, Rick Knight wrote: > Kern Sibbald wrote: > >Hello, > > > >If you are able to reproduce this easily, could you turn on level 100 by > >putting -d100 on the command line when you start it, then capture the > > output. This may help me understand what is going on. > > > >I've tried everything I can to duplicate this, but all my tests run fine. > > > >Hmmm. Normally, it wouldn't be the OS that is causing problems, but I'm > > open to almost any suggestion -- the goal being to fix it ... > > > >On Tuesday 13 December 2005 22:07, James Peverill wrote: > >>I am also getting this error since upgrading to 1.38 from the 1.36 > >>branch... been trying to figure out why for a few days now. > >> > >>I am also running slackware 10... I wonder if this is related. Not > >>running 2.4 kernel though, running 2.6.11.12 right now. I am backing up > >>to disk, not tapes. Manually scheduled jobs run just fine, and > >>automatic jobs run sometimes (the first job)... but most of the time > >>they block indefinitely with "waiting to reserve a device". > >> > >>I figured I had just screwed up a configuration file somehow, but I have > >>tweaked my configuration files with no improvement. Maybe this is tied > >>to a package in slackware 10? Any other users running slackware 10 with > >>1.38 successfully? > >> > >>james peverill > >> > >>Richard W. Knight wrote: > >>>Kern Sibbald wrote: > On Tuesday 13 December 2005 13:43, Rick Knight wrote: > >Kern Sibbald wrote: > >>Hello, > >> > >>On Monday 12 December 2005 19:35, Richard W. Knight wrote: > >>>Hi all, > >>> > >>>A couple of weeks ago I upgraded my Bacula installation from 1.34 to > >>>1.38.1. I made a few other changes at the same time and everything > >>>has > >>>been working well since. Yesterday I decided to upgrade to 1.38.3. I > >>>built from source using the same configuration options that I used > >>> to build 1.38.1 with the addition of "--with-python". The build > >>> went OK, > >>>no errors. I stopped 1.38.1 and started 1.38.3. Everything seemed > >>>to be > >>>OK. I ran a couple of small test backups and there were no errors > >>>so I > >>>assumed that the upgrade went fine. Last night the scheduled > >>>backup ran > >>>and after the first job, instead of continuing on to the next job, > >>>I got > >>>this message ... > >>> > >>>12-Dec 07:29 MyJob-SMB-sd: Job MyJob.2005-12-12_01.05.01 waiting to > >>>reserve a device. > >>> > >>>This morning, when I first saw this message, I just did a mount from > >>>bconsole and the job continued, I have bacula configured for 6 > >>>jobs per > >>>media and the tape wasn't full so the job should have just started > >>>as it > >>>hallways has. Now it's time to backup up the catalog, to file, and > >>>I'm > >>>getting the same message. > >>> > >>>I have an HP DDS2 drive, no changer, running on Slackware 10. I'm > >>>using > >>>the same conf files the worked fine on 1.38.1. Can anyone tell me > >>> why I'm now getting this error message? > >> > >>You are not by any chance running on a 2.4 kernel with /lib/tls? > >> > >>Could you send me your bacula-dir.conf and bacula-sd.conf along > >>with the > >>job report that shows the jobs blocking? > >> > >>>Thanks, > >>>RickKnight > >>> > >>> > >>>--- > >>>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=7637&alloc_id=16865&op=click > >>> ___ > >>>Bacula-users mailing list > >>>Bacula-users@lists.sourceforge.net > >>>https://lists.sourceforge.net/lists/listinfo/bacula-users > > > >Thanks Kern, > > > >I am running a 2.4.26 kernel, but I don't know about /lib/tls. I don't > >have a /lib/tls on my system. How can I tell? > > Try: > > ls -l /lib/tls > > If it exists, then that could explain why it *appears* that Bacula is > not seeing some pthread broadcasts that would allow it to continue. > This is a bit of a long shot, but at this point, I need to consider > all possibilities ... > > In the mean time, I'll take a careful look at your config. Perhaps I > have missed something important that I can add to my test cases. All > my tests here succeeded perfectly ... > > By the way, getting the message that a job is waiting to reserve a > drive is not in itself bad. This happens in my test case. However, at > some point when the drive is available, the job should continue as it > does in my test case. > >
Re: [Bacula-users] Backup to file OK, but restore fails
On 2005-12-14 17:55, Attila Fülöp wrote: > Jonas Mixter wrote: > >> On 2005-12-14 10:29, Jonas Mixter wrote: >> >> >>> Hi! >>> Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that >>> comes with Debian sarge) to 1.38.2. >>> I have one machine running the director (with a connected tape-station) >>> and one machine with a lot of disks and a storage daemon (this should >>> later on be placed elsewhere, hence the separation of the daemons). The >>> director is named merry-dir and the storage daemon is named pippin-sd. >>> Before the upgrade I could backup for example the catalog to a file >>> storage on the machine with just the sd. I could also restore files >>> from >>> the backup. >>> >>> After the upgrade, I could backup just as before. The job exits OK >>> and I >>> could see that the files that hold the backup are growing. But I cannot >>> restore... >>> When running a restore I could mark the files I want and the jobs is >>> visible underneath "Running Jobs" in bconsole. The job is "waiting for >>> Client merry-fd to connect to Storage File" and after a while the job >>> times out. >>> >>> Here's the output when trying to restore a file to merry-fd (running on >>> the same server as the director): >>> >>> 14-Dec 09:03 merry-dir: Start Restore Job >>> RestoreFiles.2005-12-14_09.03.22 >>> 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: >>> Authorization key rejected by Storage daemon. >>> Please see >>> http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors >>> for help. >>> 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: >>> Failed to authenticate Storage daemon. >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: >>> Socket error on Storage command: ERR=No data available >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula >>> 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 >>> JobId: 512 >>> Job:RestoreFiles.2005-12-14_09.03.22 >>> Client: merry-fd >>> Start time: 14-Dec-2005 09:03:25 >>> End time: 14-Dec-2005 09:13:30 >>> Files Expected: 1 >>> Files Restored: 0 >>> Bytes Restored: 0 >>> Rate: 0.0 KB/s >>> FD Errors: 0 >>> FD termination status: SD termination status: Waiting on FD >>> Termination:*** Restore Error *** >>> >>> 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula >>> 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 >>> JobId: 512 >>> Job:RestoreFiles.2005-12-14_09.03.22 >>> Client: merry-fd >>> Start time: 14-Dec-2005 09:03:25 >>> End time: 14-Dec-2005 09:13:30 >>> Files Expected: 1 >>> Files Restored: 0 >>> Bytes Restored: 0 >>> Rate: 0.0 KB/s >>> FD Errors: 1 >>> FD termination status: SD termination status: Waiting on FD >>> Termination:*** Restore Error *** >>> >>> >>> Note the job fails twice... Is that really correct? >>> I could backup and restore without any problems to a local file on the >>> server running the director (tested only once though). >>> I could check the status of the storage daemon from the director, and >>> also do backups so the passwords should be OK, right? >>> >>> I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job >>> RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but >>> don't >>> really know the meaning. In a previous post i saw that Kern asked about >>> the /lib/tls-directory when presented with this error. I do have the >>> /lib/tls, but start all my daemons with the startscript that bacula >>> provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is >>> exported. (If running kernel 2.4! is the /lib/tls a problem with kernel >>> 2.6 too?) >>> I use kernel 2.4.27 on the server with the director, and 2.6.8 on >>> the sd >>> server. The /lib/tls is there on both the servers which are running >>> Debian sarge. >>> Should really "merry-sd" be involved? I've selected the storage on >>> pippin-sd when running "restore" in bconsole. Could it be that the >>> director/fd is trying to restore from the wrong sd? Shouldn't the job >>> then fail with a "volume not found" error, ask me to mount a volume, or >>> similiar? >>> >>> If I try to cancel a job, the director from time to time complains it >>> cannot the job even though it's listed under running jobs. >>> *cancel >>> Automatically selected Job: JobId=515 >>> Job=RestoreFiles.2005-12-14_10.18.20 >>> Confirm cancel (yes/no): yes >>> 3902 Job RestoreFiles.2005-12-14_10.18.20 not found. >>> >>> The job is marked "has been canceled" in bconsole anyway. I don't >>> remember if I got errors like this before the upgrade. >>> >>> What could be wrong in my setup? I've been banging my head to the wall >>> for quite a few hours now and I'm out of ideas. >>> >>> Best regards, Jonas Mixter
Re: [Bacula-users] Error "waiting to reserve a device" after upgrade to 1.38.3
Hello, Please copy the list ... On Wednesday 14 December 2005 04:12, Rick Knight wrote: > Kern Sibbald wrote: > >Hello, > > > >If you are able to reproduce this easily, could you turn on level 100 by > >putting -d100 on the command line when you start it, then capture the > > output. This may help me understand what is going on. > > > >I've tried everything I can to duplicate this, but all my tests run fine. > > > >Hmmm. Normally, it wouldn't be the OS that is causing problems, but I'm > > open to almost any suggestion -- the goal being to fix it ... > > > >On Tuesday 13 December 2005 22:07, James Peverill wrote: > >>I am also getting this error since upgrading to 1.38 from the 1.36 > >>branch... been trying to figure out why for a few days now. > >> > >>I am also running slackware 10... I wonder if this is related. Not > >>running 2.4 kernel though, running 2.6.11.12 right now. I am backing up > >>to disk, not tapes. Manually scheduled jobs run just fine, and > >>automatic jobs run sometimes (the first job)... but most of the time > >>they block indefinitely with "waiting to reserve a device". > >> > >>I figured I had just screwed up a configuration file somehow, but I have > >>tweaked my configuration files with no improvement. Maybe this is tied > >>to a package in slackware 10? Any other users running slackware 10 with > >>1.38 successfully? > >> > >>james peverill > >> > >>Richard W. Knight wrote: > >>>Kern Sibbald wrote: > On Tuesday 13 December 2005 13:43, Rick Knight wrote: > >Kern Sibbald wrote: > >>Hello, > >> > >>On Monday 12 December 2005 19:35, Richard W. Knight wrote: > >>>Hi all, > >>> > >>>A couple of weeks ago I upgraded my Bacula installation from 1.34 to > >>>1.38.1. I made a few other changes at the same time and everything > >>>has > >>>been working well since. Yesterday I decided to upgrade to 1.38.3. I > >>>built from source using the same configuration options that I used > >>> to build 1.38.1 with the addition of "--with-python". The build > >>> went OK, > >>>no errors. I stopped 1.38.1 and started 1.38.3. Everything seemed > >>>to be > >>>OK. I ran a couple of small test backups and there were no errors > >>>so I > >>>assumed that the upgrade went fine. Last night the scheduled > >>>backup ran > >>>and after the first job, instead of continuing on to the next job, > >>>I got > >>>this message ... > >>> > >>>12-Dec 07:29 MyJob-SMB-sd: Job MyJob.2005-12-12_01.05.01 waiting to > >>>reserve a device. > >>> > >>>This morning, when I first saw this message, I just did a mount from > >>>bconsole and the job continued, I have bacula configured for 6 > >>>jobs per > >>>media and the tape wasn't full so the job should have just started > >>>as it > >>>hallways has. Now it's time to backup up the catalog, to file, and > >>>I'm > >>>getting the same message. > >>> > >>>I have an HP DDS2 drive, no changer, running on Slackware 10. I'm > >>>using > >>>the same conf files the worked fine on 1.38.1. Can anyone tell me > >>> why I'm now getting this error message? > >> > >>You are not by any chance running on a 2.4 kernel with /lib/tls? > >> > >>Could you send me your bacula-dir.conf and bacula-sd.conf along > >>with the > >>job report that shows the jobs blocking? > >> > >>>Thanks, > >>>RickKnight > >>> > >>> > >>>--- > >>>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=7637&alloc_id=16865&op=click > >>> ___ > >>>Bacula-users mailing list > >>>Bacula-users@lists.sourceforge.net > >>>https://lists.sourceforge.net/lists/listinfo/bacula-users > > > >Thanks Kern, > > > >I am running a 2.4.26 kernel, but I don't know about /lib/tls. I don't > >have a /lib/tls on my system. How can I tell? > > Try: > > ls -l /lib/tls > > If it exists, then that could explain why it *appears* that Bacula is > not seeing some pthread broadcasts that would allow it to continue. > This is a bit of a long shot, but at this point, I need to consider > all possibilities ... > > In the mean time, I'll take a careful look at your config. Perhaps I > have missed something important that I can add to my test cases. All > my tests here succeeded perfectly ... > > By the way, getting the message that a job is waiting to reserve a > drive is not in itself bad. This happens in my test case. However, at > some point when the drive is available, the job should continue as it >
Re: [Bacula-users] Cycling a USB drive - confused...
Hi Ferdinando > I have done something silmilar with tape > magazines. I use normal backups on disks, with > Full on sunday and Incr on weekdays. > > On saturday I schedule a Full copy on > tapes (2 drives on a 19 slots changer). We keep > 3 magazines wich are rotated, > each containing some volumes of the 4 pools onto > which we make the backups. > Volumes expire after 24 days. > > On monday we can verify if all the backups > were done correctly (eventually we have all the > week to integrate failed ones). > > When we want to change the magazine > we use a script (through WEBMIN) that unmount > and unload the drives, then > we change the magazine and run another script > (really is the same with > another argument) that disables the volumes in > the previous mag and enable the new ones. > > I will attach the script, but it is > not general and really not commented. Also it is > working with 1.36.3 and messages are in Italian. > > If someone in the list is interested I could try > to make it more general and usable. > > Ferdinando Pasqualetti > G.T.Dati srl > Tel. 0557310862 - 3356172731 - Fax 055720143 I think I can use this, no problem with the Italian messages. Another job for the holiday period! Thanks and regards, -- Richard (MQ) --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] clientrunbeforejob not working
Hi, Computer King schrieb: I am tring to run the following 2 commands with bacula but them do not seem to work when i run them on the local machine it works fine however these commands must be run as root how can i get this to work with zimbra. ClientRunBeforeJob = /opt/zimbra/bin/zimbra stop ClientRunAfterJob = /opt/zimbra/bin/zimbra start Tried su, sudo or setting the file permissions to... erm... SETUID? Arno --- 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=7637&alloc_id=16865&op=click ___ 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 --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Sparc testing
Hello,
Is there someone on this list with a Sparc or any other machine that is
running a Storage daemon other than Linux with kernel 2.6 or FreeBSD where I
am able to test myself? (e.g. Sparc, Linux with kernel 2.4, OpenBSD,
IBM, ...)
I would like you to run about 10 minutes of tests with and without a tape in a
drive (or autochanger). If so, please let me know, and I'll send you the
program and the commands to run.
--
Best regards,
Kern
(">
/\
V_V
---
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=7637&alloc_id=16865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] clientrunbeforejob not working
I am tring to run the following 2 commands with bacula but them do not seem to work when i run them on the local machine it works fine however these commands must be run as root how can i get this to work with zimbra. ClientRunBeforeJob = /opt/zimbra/bin/zimbra stop ClientRunAfterJob = /opt/zimbra/bin/zimbra start --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Feature Request
I haven't seen it in the documentation in my travels, so i'll assume this isn't possible at the moment. Kern, i'm backing up 49 servers at the moment and occasionally one of their network cards will get the wrong duplex setting, so the backup will go at 50K/sec. Can we have a setting which cancels a job if it's not achieving a decent throughput? My problem is, if the 10th server screws up, I check it in the morning and that one server stops the rest from having a succesful backup. Cheers and keep up the great work! Beren *** Mail FROM London Borough of Harrow: Unencrypted electronic mail is not secure and may not be authentic, in whole or in part. You are advised to check directly with the sender before acting upon any e-mail received. The information contained in this message and any attachments is confidential and is intended for receipt by the above named addressee(s) only. If you have otherwise encountered this message please notify its originator via +44(0)20 8863 5611 at LONDON BOROUGH OF HARROW. The unauthorised use, disclosure, copying or alteration of this message is strictly forbidden. The views expressed within this message are those of the individual sender and not necessarily those of Harrow Council. Mail TO London Borough of Harrow: London Borough of Harrow monitors all electronic mail it receives for Policy compliance and to protect its systems including anti-spam and anti-virus measures. Electronic mail does not guarantee delivery, nor notification of non-delivery. It is suggested you contact your intended recipient(s) by other means should confirmation of receipt be important. *** --- 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_idv37&alloc_id865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Error "waiting to reserve a device" after upgrade to 1.38.3
Kern Sibbald wrote: Hello, If you are able to reproduce this easily, could you turn on level 100 by putting -d100 on the command line when you start it, then capture the output. This may help me understand what is going on. I've tried everything I can to duplicate this, but all my tests run fine. Hmmm. Normally, it wouldn't be the OS that is causing problems, but I'm open to almost any suggestion -- the goal being to fix it ... On Tuesday 13 December 2005 22:07, James Peverill wrote: Kern, I modified the bacula startup script to start with -d100. Attached is the output. Also, I noticed that when the first job ends, Knight-Linux-SMB, bconsole/status/dir shows the drive as not mounted. When I mount the drive, bacula claims the drive is already mounted, but then the job continues. My runafter script has not run at this point, but when it does, it also ejects the tape so I don't think that it's causing this problem. Also, looking at the attached file, I can see a couple of python errors "pythonlib.c:94 No script dir". Could this be the problem? Thank again, RickKnight [EMAIL PROTECTED]:/etc/bacula# ./bacula start Starting the Bacula Storage daemon bacula-sd: stored_conf.c:613 Inserting device res: DAT-8 Starting the Bacula File daemon knight-linux-SMB-sd: jcr.c:113 read_last_jobs seek to 188 knight-linux-SMB-sd: jcr.c:120 Read num_items=10 knight-linux-SMB-sd: pythonlib.c:94 No script dir. prog=SDStartUp knight-linux-SMB-sd: stored.c:451 calling init_dev /tmp/bacula_cat knight-linux-SMB-sd: dev.c:242 init_dev: tape=0 dev_name=/tmp/bacula_cat knight-linux-SMB-sd: stored.c:453 SD init done /tmp/bacula_cat knight-linux-SMB-sd: stored.c:451 calling init_dev /dev/nst0 knight-linux-SMB-sd: dev.c:242 init_dev: tape=2 dev_name=/dev/nst0 knight-linux-SMB-sd: stored.c:453 SD init done /dev/nst0 knight-linux-SMB-sd: bnet_server.c:83 Addresses host[ipv4:0.0.0.0:9103] knight-linux-SMB-fd: jcr.c:113 read_last_jobs seek to 188 knight-linux-SMB-fd: jcr.c:120 Read num_items=10 knight-linux-SMB-fd: pythonlib.c:94 No script dir. prog=FDStartUp knight-linux-SMB-fd: filed.c:238 filed: listening on port 9102 knight-linux-SMB-fd: bnet_server.c:83 Addresses host[ipv4:0.0.0.0:9102] Starting the Bacula Director daemon bacula-dir: dird.c:128 Debug level = 100 bacula-dir: mysql.c:81 db_open first time bacula-dir: mysql.c:142 mysql_init done bacula-dir: mysql.c:162 mysql_real_connect done bacula-dir: mysql.c:164 db_user=bacula db_name=bacula db_password=passwd_deleted knight-linux-SMB-dir: jcr.c:113 read_last_jobs seek to 188
Re: [Bacula-users] Retoring files very slow
On Wednesday 14 December 2005 16:28, Bruno Léon wrote:
> Hi all,
>
> I'm using bacula for 2 months now and I'm doing some restore tests now
> that the volume of data is quite big.
> Actually, I'd like to restore a file but this takes an extremely long
> time, more than 2 hours to read the catalog.
> The catalog is 500Mo and the server is only backup server, 1.3Ghz PIII
> with 256 Mb or RAM.
>
> Do you think that this duration is normal or not ?
Yes, that is abnormal. Please read the Catalog Maintenance chapter in the
development manual. I just added a small section on database performance, in
particular checking that you have the right indexes ...
>
> Thanks in advance.
--
Best regards,
Kern
(">
/\
V_V
---
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_idv37&alloc_id865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Re: Segmentation violation
On Wednesday 14 December 2005 15:59, Roger Kvam wrote: > Now I`m really embarrassed, actually ,FreeBSD is protecting itself > against runaway processes by allowing maximum memory size for one > process to 512MB, I had to increase the maximum to 2 gig by adding > kern.maxdsiz=2147483648 to /boot/loader.conf. witch is my physical > amount of memory, but maxed out that to, so I had tried another job who > only used 870MB RAM. Maybe there should be a warning in the > documentation about this problem? I think my job would have required at > least 2.3GB memory :D I will divide it now ;) Well, there is nothing to be embarrassed about. I had no idea that FreeBSD limited the maximum memory. I agree this should be in the doc, and I will add it. > > But the error messages from bacula was far out :p > "Query failed: SELECT MediaType > > >>> FROM JobMedia,Media WHERE JobMedia.JobId=2 AND > >>> JobMedia.MediaId=Media.MediaId: ERR=Lost connection to MySQL server > >>> during query" > > Roger > > Roger Kvam wrote: > > I have tried to run several restorejobs, and every job who is beneth 516 > > MB i memory consumption is a sucess, every job above 516 crash whith > > same error message. MySQL is installed same way as when running bacula > > 1.36. > > > > Bacula did crash earlyer when consuming more than fysical mermory, but > > the box was upgraded to 2GB prior to 1.36, and the versions prior to > > 1.36 and 1.36 worked fine. > > > > Kern Sibbald wrote: > >> Hello, > >> > >> I repeat what I said the last time: this looks like a MySQL problem. > >> Given the new information you have presented, and the fact that you > >> are having problems when working with almost 9 million files backed up > >> (a lot), I would suspect that the problem is with your MySQL > >> configuration. If you used the standard MySQL installation, it is > >> probably not configured for such large databases. > >> > >> See below for more ... > >> > >> On Monday 12 December 2005 10:52, Roger Kvam wrote: > >>> I Tried to upgrade MySQL to 5. and recomile bacula (yes I did a make > >>> distclean), but same error, have now installed MySQL 4.1.15 again, and > >>> recompiled bacula again, deleted the database and deleted the tapes so > >>> I could start fresh. > >>> > >>> When trying to restore from a large job, bacula crashes, > >> > >> From the message you show below, it looks to me like MySQL crashed or > >> at least disconnected. > >> > >>> when trying to restore from a small job everything works fine. Does > >>> bacula has > >>> shortcomings when comes to enterprise systems? > >> > >> Bacula is not what I would call an enterprise system, so one could say > >> it has that kind of shortcoming. In your case, you are dealing with > >> a quite large backup set, and it looks like you have not adapted MySQL > >> to handle it appropriatedly. > >> > >> Perhaps some of the other users on the list can help you. If I am not > >> mistaken, the Bacula document also mentions some steps you might take > >> for large databases, but your best bet is to look at the MySQL > >> documentation. > >> > >>> Now I have run several backups of several servers and everything works > >>> fine, until I wanted to do a restore from the biggest job, > >>> > >>> +---+---+---+-+-+-- > >>> > >>> > >>> --+---+ > >>> > >>> | JobId | Level | JobFiles | JobBytes| StartTime | > >>> > >>> VolumeName | StartFile | > >>> +---+---+---+-+-+-- > >>> > >>> > >>> --+---+ > >>> > >>> | 2 | F | 8,759,302 | 821,505,476,167 | 2005-12-09 16:27:02 | > >>> > >>> 09L2 | 0 | > >>> +---+---+---+-+-+-- > >>> > >>> > >>> --+---+ You have selected the following JobId: 2 > >>> > >>> Building directory tree for JobId 2 ... Query failed: SELECT MediaType > >>> FROM JobMedia,Media WHERE JobMedia.JobId=2 AND > >>> JobMedia.MediaId=Media.MediaId: ERR=Lost connection to MySQL server > >>> during query > >>> > >>> There were no files inserted into the tree, so file selection > >>> is not possible.Most likely your retention policy pruned the files > >>> > >>> Do you want to restore all the files? (yes|no): > >>> ¤¤ > >>> > >>> When trying to restore from a smaller job, everything works fine: > >>> > >>> +---+---+--++-+ > >>> > >>> > >>> +---+ > >>> > >>> | JobId | Level | JobFiles | JobBytes | StartTime | > >>> > >>> VolumeName | StartFile | > >>> +---+---+--++-+ > >>> > >>> > >>> +---+ > >>> > >>> | 3 | F | 833,832 | 52,358,007,603 | 2005-12-09 16:29:06 | > >>> > >>> 09L2 |29 | > >>> > >>> | 4 | I |1,434 |322,179,540 | 2005-12-
Re: [Bacula-users] Backup to file OK, but restore fails
Jonas Mixter wrote: On 2005-12-14 10:29, Jonas Mixter wrote: Hi! Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that comes with Debian sarge) to 1.38.2. I have one machine running the director (with a connected tape-station) and one machine with a lot of disks and a storage daemon (this should later on be placed elsewhere, hence the separation of the daemons). The director is named merry-dir and the storage daemon is named pippin-sd. Before the upgrade I could backup for example the catalog to a file storage on the machine with just the sd. I could also restore files from the backup. After the upgrade, I could backup just as before. The job exits OK and I could see that the files that hold the backup are growing. But I cannot restore... When running a restore I could mark the files I want and the jobs is visible underneath "Running Jobs" in bconsole. The job is "waiting for Client merry-fd to connect to Storage File" and after a while the job times out. Here's the output when trying to restore a file to merry-fd (running on the same server as the director): 14-Dec 09:03 merry-dir: Start Restore Job RestoreFiles.2005-12-14_09.03.22 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Failed to authenticate Storage daemon. 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: Socket error on Storage command: ERR=No data available 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 0 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 1 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** Note the job fails twice... Is that really correct? I could backup and restore without any problems to a local file on the server running the director (tested only once though). I could check the status of the storage daemon from the director, and also do backups so the passwords should be OK, right? I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but don't really know the meaning. In a previous post i saw that Kern asked about the /lib/tls-directory when presented with this error. I do have the /lib/tls, but start all my daemons with the startscript that bacula provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is exported. (If running kernel 2.4! is the /lib/tls a problem with kernel 2.6 too?) I use kernel 2.4.27 on the server with the director, and 2.6.8 on the sd server. The /lib/tls is there on both the servers which are running Debian sarge. Should really "merry-sd" be involved? I've selected the storage on pippin-sd when running "restore" in bconsole. Could it be that the director/fd is trying to restore from the wrong sd? Shouldn't the job then fail with a "volume not found" error, ask me to mount a volume, or similiar? If I try to cancel a job, the director from time to time complains it cannot the job even though it's listed under running jobs. *cancel Automatically selected Job: JobId=515 Job=RestoreFiles.2005-12-14_10.18.20 Confirm cancel (yes/no): yes 3902 Job RestoreFiles.2005-12-14_10.18.20 not found. The job is marked "has been canceled" in bconsole anyway. I don't remember if I got errors like this before the upgrade. What could be wrong in my setup? I've been banging my head to the wall for quite a few hours now and I'm out of ideas. Best regards, Jonas Mixter Hi! I've doing some more tests today. There seems to be no problems at all restoring from backups stored on tape. If I copy the entire backupfile from the server with running only the sd, to the server running the bacula-dir I could do a restore files without any problems. Does anyone have any suggestions where continue trouble shooting? Best regards, Jonas Mixter Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-man
Re: [Bacula-users] Re: Segmentation violation
Heigl Florian - Munich-MR - external wrote: Hi Roger, Now I`m really embarrassed, actually ,FreeBSD is protecting itself against runaway processes by allowing maximum memory size for one process to 512MB, I had to increase the maximum to 2 gig by adding kern.maxdsiz=2147483648 to /boot/loader.conf. witch is my physical amount of memory, but maxed out that to, so I had tried another job who only used 870MB RAM. Maybe there should be a warning in the documentation about this problem? I think my job would have required at least 2.3GB memory :D I will divide it now ;) You should still keep maxdsiz below Your physical memory or You'll end up either trashing Your paging space or getting random processes killed by the OS. Do some sizing of the mysql database configuration and let it know it can't allocate more than 1GB physical memory and set maxdsiz to that. Well this all depends on your hardware. On 32 bit systems physical address space is limited to 4G (32 bit pointers), out of which the PCI address space takes the half. Additional address space is needed for mmaping and shared libraries. This means that on 32 bit platforms your limit per process is below 2G. This is a physical limitation. The 512MB limit is documented and explained in the manual. I had to tweak a Xeon server with FreeBSD 4.8 to use 1.8G per process. and found the information quite quickly googling around. (Sorry I've no bookmark right now.) I have no problems at all with a single process consuming more than 2Gig on an AMD Opteron system running FreeBSD 5.4 without tweaking anything. Attila Florian --- 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_idv37&alloc_id865&opÌk ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users --- 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_idv37&alloc_id865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Backup to file OK, but restore fails
On 2005-12-14 10:29, Jonas Mixter wrote: >Hi! >Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that >comes with Debian sarge) to 1.38.2. >I have one machine running the director (with a connected tape-station) >and one machine with a lot of disks and a storage daemon (this should >later on be placed elsewhere, hence the separation of the daemons). The >director is named merry-dir and the storage daemon is named pippin-sd. >Before the upgrade I could backup for example the catalog to a file >storage on the machine with just the sd. I could also restore files from >the backup. > >After the upgrade, I could backup just as before. The job exits OK and I >could see that the files that hold the backup are growing. But I cannot >restore... >When running a restore I could mark the files I want and the jobs is >visible underneath "Running Jobs" in bconsole. The job is "waiting for >Client merry-fd to connect to Storage File" and after a while the job >times out. > >Here's the output when trying to restore a file to merry-fd (running on >the same server as the director): > >14-Dec 09:03 merry-dir: Start Restore Job RestoreFiles.2005-12-14_09.03.22 >14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: >Authorization key rejected by Storage daemon. >Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors >for help. >14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: >Failed to authenticate Storage daemon. >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: >Socket error on Storage command: ERR=No data available >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula >1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > JobId: 512 > Job:RestoreFiles.2005-12-14_09.03.22 > Client: merry-fd > Start time: 14-Dec-2005 09:03:25 > End time: 14-Dec-2005 09:13:30 > Files Expected: 1 > Files Restored: 0 > Bytes Restored: 0 > Rate: 0.0 KB/s > FD Errors: 0 > FD termination status: > SD termination status: Waiting on FD > Termination:*** Restore Error *** > >14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula >1.38.2 (20Nov05): 14-Dec-2005 09:13:30 > JobId: 512 > Job:RestoreFiles.2005-12-14_09.03.22 > Client: merry-fd > Start time: 14-Dec-2005 09:03:25 > End time: 14-Dec-2005 09:13:30 > Files Expected: 1 > Files Restored: 0 > Bytes Restored: 0 > Rate: 0.0 KB/s > FD Errors: 1 > FD termination status: > SD termination status: Waiting on FD > Termination:*** Restore Error *** > > >Note the job fails twice... Is that really correct? >I could backup and restore without any problems to a local file on the >server running the director (tested only once though). >I could check the status of the storage daemon from the director, and >also do backups so the passwords should be OK, right? > >I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job >RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but don't >really know the meaning. In a previous post i saw that Kern asked about >the /lib/tls-directory when presented with this error. I do have the >/lib/tls, but start all my daemons with the startscript that bacula >provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is >exported. (If running kernel 2.4! is the /lib/tls a problem with kernel >2.6 too?) >I use kernel 2.4.27 on the server with the director, and 2.6.8 on the sd >server. The /lib/tls is there on both the servers which are running >Debian sarge. >Should really "merry-sd" be involved? I've selected the storage on >pippin-sd when running "restore" in bconsole. Could it be that the >director/fd is trying to restore from the wrong sd? Shouldn't the job >then fail with a "volume not found" error, ask me to mount a volume, or >similiar? > >If I try to cancel a job, the director from time to time complains it >cannot the job even though it's listed under running jobs. >*cancel >Automatically selected Job: JobId=515 Job=RestoreFiles.2005-12-14_10.18.20 >Confirm cancel (yes/no): yes >3902 Job RestoreFiles.2005-12-14_10.18.20 not found. > >The job is marked "has been canceled" in bconsole anyway. I don't >remember if I got errors like this before the upgrade. > >What could be wrong in my setup? I've been banging my head to the wall >for quite a few hours now and I'm out of ideas. > >Best regards, Jonas Mixter > Hi! I've doing some more tests today. There seems to be no problems at all restoring from backups stored on tape. If I copy the entire backupfile from the server with running only the sd, to the server running the bacula-dir I could do a restore files without any problems. Does anyone have any suggestions where continue trouble
Re: [Bacula-users] Retoring files very slow
Please always copy the list! Personal support is restricted to people who pay me or who are my father-in-law. Thanks. On Wednesday 14 December 2005 16:51, Bruno Léon wrote: > Well, I only have an Index on FilenameID at the moment. > May be those new indexes come with bacula 1.38? I'm using 1.36.2 which > comes as a Debian package. I'm using the Debian/Sarge package, too. And it's true that the indexes are currently not created automatically. See my submitted bug report: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=335809 Create the indices and you will see a performance boost that will make you happy. Regards Christoph -- ~ ~ ".signature" [Modified] 2 lines --100%--2,41 All --- 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_idv37&alloc_id865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Retoring files very slow
On Wed, Dec 14, 2005 at 04:56:08PM +0100, Alejandro Alfonso wrote: > Witch database do you use? Try to dump and then restore the databasefile > to rebuild the Indexes As a quick sidenote, it's a lot easier and safer to use the optimize command to rebuild indexes on tables. http://dev.mysql.com/doc/refman/4.1/en/optimize-table.html -- Frank Sweetser fs at wpi.edu | For every problem, there is a solution that WPI Network Engineer | is simple, elegant, and wrong. - HL Mencken GPG fingerprint = 6174 1257 129E 0D21 D8D4 E8A3 8E39 29E3 E2E8 8CEC --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Retoring files very slow
Build a "restore enviroment" takes me about five minutes in a: Gentoo Linux 2.6.11 1) Bacula 1.38.0 2) Mysql 4.1 3) ~2.000.000 files in the job 4) 1'5Gb /var/lib/mysql Hardware... Pentium II (yes!!) 128Mb RAM (yes!!!) harddisk 8Gb SCSI2 Witch database do you use? Try to dump and then restore the databasefile to rebuild the Indexes Best regards Hi all, I'm using bacula for 2 months now and I'm doing some restore tests now that the volume of data is quite big. Actually, I'd like to restore a file but this takes an extremely long time, more than 2 hours to read the catalog. The catalog is 500Mo and the server is only backup server, 1.3Ghz PIII with 256 Mb or RAM. Do you think that this duration is normal or not ? Thanks in advance. -- Alejandro Alfonso Fernández Dpto de Sistemas. Área Corporativa [EMAIL PROTECTED] http://www.telecyl.com/ Proción 7, Portales 1-2 Edificio América II 28023 Madrid Tfn: 91 452 18 00 - Fax: 91 452 18 08 Juan García Hortelano, 43 Edificio Telecyl 47014 Valladolid Tfn: 983 428 200 - Fax: 983 428 223
Re: [Bacula-users] Retoring files very slow
On Wednesday 14 December 2005 16:28, Bruno Léon wrote: > I'm using bacula for 2 months now and I'm doing some restore tests now > that the volume of data is quite big. > Actually, I'd like to restore a file but this takes an extremely long > time, more than 2 hours to read the catalog. > The catalog is 500Mo and the server is only backup server, 1.3Ghz PIII > with 256 Mb or RAM. > > Do you think that this duration is normal or not ? I don't think so. I had similar problems when the indices on my MySQL database haven't been created properly. Check that you have the following indices created on the `File` table: # INDEX (JobId), # INDEX (PathId), # INDEX (FilenameId), # INDEX (JobId, PathId, FilenameId) Currently it takes me a maximum of 30 seconds to build the directory tree. Without having these indices it took 5-10 minutes here. Regards Christoph -- ~ ~ ".signature" [Modified] 2 lines --100%--2,41 All --- 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_idv37&alloc_id865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Retoring files very slow
Hi all, I'm using bacula for 2 months now and I'm doing some restore tests now that the volume of data is quite big. Actually, I'd like to restore a file but this takes an extremely long time, more than 2 hours to read the catalog. The catalog is 500Mo and the server is only backup server, 1.3Ghz PIII with 256 Mb or RAM. Do you think that this duration is normal or not ? Thanks in advance. -- Bruno LEON begin:vcard fn;quoted-printable:Bruno L=C3=A9on n;quoted-printable:L=C3=A9on;Bruno org:Agreentech adr:Parc d'affaires Metropolis CS20 825;;14 E Rue Patis Tatelin;Rennes Cedex 7;;35708;FRANCE email;internet:[EMAIL PROTECTED] title;quoted-printable:Administrateur Syst=C3=A8mes & R=C3=A9seaux x-mozilla-html:TRUE url:http://www.agreentech.com version:2.1 end:vcard
RE: [Bacula-users] Re: Segmentation violation
Hi Roger, > Now I`m really embarrassed, actually ,FreeBSD is protecting itself > against runaway processes by allowing maximum memory size for one > process to 512MB, I had to increase the maximum to 2 gig by adding > kern.maxdsiz=2147483648 to /boot/loader.conf. witch is my physical > amount of memory, but maxed out that to, so I had tried > another job who > only used 870MB RAM. Maybe there should be a warning in the > documentation about this problem? I think my job would have > required at > least 2.3GB memory :D I will divide it now ;) You should still keep maxdsiz below Your physical memory or You'll end up either trashing Your paging space or getting random processes killed by the OS. Do some sizing of the mysql database configuration and let it know it can't allocate more than 1GB physical memory and set maxdsiz to that. Florian --- 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_idv37&alloc_id865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Re: Segmentation violation
Now I`m really embarrassed, actually ,FreeBSD is protecting itself against runaway processes by allowing maximum memory size for one process to 512MB, I had to increase the maximum to 2 gig by adding kern.maxdsiz=2147483648 to /boot/loader.conf. witch is my physical amount of memory, but maxed out that to, so I had tried another job who only used 870MB RAM. Maybe there should be a warning in the documentation about this problem? I think my job would have required at least 2.3GB memory :D I will divide it now ;) But the error messages from bacula was far out :p "Query failed: SELECT MediaType >>> FROM JobMedia,Media WHERE JobMedia.JobId=2 AND >>> JobMedia.MediaId=Media.MediaId: ERR=Lost connection to MySQL server >>> during query" Roger Roger Kvam wrote: I have tried to run several restorejobs, and every job who is beneth 516 MB i memory consumption is a sucess, every job above 516 crash whith same error message. MySQL is installed same way as when running bacula 1.36. Bacula did crash earlyer when consuming more than fysical mermory, but the box was upgraded to 2GB prior to 1.36, and the versions prior to 1.36 and 1.36 worked fine. Kern Sibbald wrote: Hello, I repeat what I said the last time: this looks like a MySQL problem. Given the new information you have presented, and the fact that you are having problems when working with almost 9 million files backed up (a lot), I would suspect that the problem is with your MySQL configuration. If you used the standard MySQL installation, it is probably not configured for such large databases. See below for more ... On Monday 12 December 2005 10:52, Roger Kvam wrote: I Tried to upgrade MySQL to 5. and recomile bacula (yes I did a make distclean), but same error, have now installed MySQL 4.1.15 again, and recompiled bacula again, deleted the database and deleted the tapes so I could start fresh. When trying to restore from a large job, bacula crashes, From the message you show below, it looks to me like MySQL crashed or at least disconnected. when trying to restore from a small job everything works fine. Does bacula has shortcomings when comes to enterprise systems? Bacula is not what I would call an enterprise system, so one could say it has that kind of shortcoming. In your case, you are dealing with a quite large backup set, and it looks like you have not adapted MySQL to handle it appropriatedly. Perhaps some of the other users on the list can help you. If I am not mistaken, the Bacula document also mentions some steps you might take for large databases, but your best bet is to look at the MySQL documentation. Now I have run several backups of several servers and everything works fine, until I wanted to do a restore from the biggest job, +---+---+---+-+-+-- --+---+ | JobId | Level | JobFiles | JobBytes| StartTime | VolumeName | StartFile | +---+---+---+-+-+-- --+---+ | 2 | F | 8,759,302 | 821,505,476,167 | 2005-12-09 16:27:02 | 09L2 | 0 | +---+---+---+-+-+-- --+---+ You have selected the following JobId: 2 Building directory tree for JobId 2 ... Query failed: SELECT MediaType FROM JobMedia,Media WHERE JobMedia.JobId=2 AND JobMedia.MediaId=Media.MediaId: ERR=Lost connection to MySQL server during query There were no files inserted into the tree, so file selection is not possible.Most likely your retention policy pruned the files Do you want to restore all the files? (yes|no): ¤¤ When trying to restore from a smaller job, everything works fine: +---+---+--++-+ +---+ | JobId | Level | JobFiles | JobBytes | StartTime | VolumeName | StartFile | +---+---+--++-+ +---+ | 3 | F | 833,832 | 52,358,007,603 | 2005-12-09 16:29:06 | 09L2 |29 | | 4 | I |1,434 |322,179,540 | 2005-12-10 00:05:01 | 01L2 | 0 | +---+---+--++-+ +---+ You have selected the following JobIds: 3,4 Building directory tree for JobId 3 ... + Building directory tree for JobId 4 ... 2 Jobs, 828,824 files inserted into the tree. Have I met any limitations in bacula? The system status was like this the moment before the crash: last pid: 34607; load averages: 1.24, 0.58, 0.25 up 3+19:19:56 08:57:34 59 processes: 6 running, 53 sleeping Mem: 572M Active, 1156M Inact, 212M Wired, 64M Cache, 112M Buf, 3008K Free Swap: 4096M Total, 500K Used, 4095M Free PID USERNAMETHR PRI NICE
Re: [Bacula-users] Re: Bacula BETA 1.38.3
Hello Arno, I've been spending a lot of time thinking about this problem lately, trying to take in all the information you have supplied. For the moment, the most promising idea is what you found with unloaded drives. By the way, I am convinced that the problems Rick Knight are seeing are totally unrelated. On Wednesday 14 December 2005 09:28, Arno Lehmann wrote: > Hello, > > some more results. I was curious, so I found the opportunity to verify > at least part of todays regular backup run. > > Arno Lehmann schrieb: > > Kern - > > > > Kern Sibbald schrieb: > >> Hmmm. I think I am going to pack it in and take a year's vacation ... > > > > don't. Yea, I won't, but I am sure getting annoyed with this problem. Please don't take that to mean I am annoyed with you, quite the contrary, you have been very patient. > > > > I was a little stupid when I tried to analyze the problem. Or rather, I > > didn't really analyze. > > > > What I did this evening: Shutdown bacula. Use mtx to load a tape into > > ONE of my TWO autoloaders (might be important later). > > Start Bacula again. > > > > Start a job going to autoloader ONE. > > > > Voilà. > > > > New results now: When Bacula encounters an empty autoloader drive at > > startup, I can't get it loading and using tapes. When the drive is > > loaded when Bacula (SD, probably) starts, everything's ok. Well, this is not at all good, but it is probably *very* important. If I remember right, you are using SUSE, can you remind me of what kernel you are on. I ask because with 2.6 kernels (actually 2.5.x) there was this major change that was discussed some time ago where if there is no tape in the drive, an open() in read/write mode waits for some period of time, then fails. Now, perhaps (very unlikely) on your system, it waits forever. Before anything else can work, Bacula needs to be able to open a tape drive correctly. On my system Fedora FC4 (kernel 2.6.14), it works fine. Perhaps at this point, I need to write a little test program that you can run for me that will report back what is going on. > > > > What I will / might / should test: > > - What happens with the empty autoloader TWO? I suspect it will remain > > blocked. > > - What happens if I umount ONE and later mount it again? > > I had to change tapes. The normal umount - change magazines - update > slots scan - mount worked. Jobs that go to the loaded volumes run, > others wait. No jobs failed until now, but up till now my original > problem situation didn't happen. > > > Tomorrow, I've got a job scheduled that should go to TWO - I'll report. > > This job is still waiting for Maximum Concurrent Jobs on the client... OK, Regards, Kern > > Arno > > > Arno > > > >> On Monday 12 December 2005 20:21, Arno Lehmann wrote: > >>> Hi, > >>> > >>> Arno Lehmann schrieb: > Hi, > > Kern Sibbald schrieb: > >>> > >>> ... > >>> > > Anyway, I would really appreciate it if you could try the 1.38.3 > > beta. Please use the released tar file rather than the CVS -- the > > HEAD CVS is now slightly behind the 1.38 branch in terms of fixes ... > > I'll download it in a few minutes... > > > Although I haven't been able to reproduce the pool problems you are > > having, I have been working on it nevertheless. When I was able to > > duplicate the two drive problem of not running jobs simultaneously on > > both drives, I took the time to review the reservation system in > > detail, and I found a few places where there *may* have been > > problems. > > I should see if the version 1.38.3beta fixes my problem on Tuesday - > that's still my jobs-to-different-pools-day. I'll report what happens. > And, of course, I've got one run of straight and usual backups before > that, although I don't think there will be any problems tomorrow. > >>> > >>> Unfortunately, I have to report something rather unexpected. The device > >>> reservation system works a little too good ;-) > >>> > >>> I started my bachup server this morning (or rather, it started itself). > >>> All tape drives Bacula itself loads were empty, magazines were filled > >>> with the correct volumes. Basically, the same situation I have 5 out of > >>> seven days for about a year now, with many versions of Bacula. > >>> > >>> Jobs were scheduled as usual and got triggered as usual, too. > >>> > >>> After that, not much progress since this morning, 8:20 (almost 12 hours > >>> > >>> ago): > 12-Dec 11:21 goblin-sd: Job DracheStd.2005-12-12_08.20.00 waiting to > reserve a device. 12-Dec 11:26 goblin-sd: Job > Goblin.2005-12-12_08.25.00 > waiting to reserve a device. 12-Dec 11:26 goblin-sd: Job > Ork.2005-12-12_08.25.03 waiting to reserve a device. 12-Dec 11:26 > goblin-sd: Job ElfSys.2005-12-12_08.25.01 waiting to reserve a device. > 12-Dec 11:26 goblin-sd: Job ElfHome.2005-12-12_08.25.02 waiting to > reserve a device. 12-Dec 15:21 go
Re: [Bacula-users] Re: Bacula BETA 1.38.3
On Monday 12 December 2005 20:21, Arno Lehmann wrote:
> Hi,
>
> Arno Lehmann schrieb:
> > Hi,
> >
> > Kern Sibbald schrieb:
>
> ...
>
> >> Anyway, I would really appreciate it if you could try the 1.38.3
> >> beta. Please use the released tar file rather than the CVS -- the
> >> HEAD CVS is now slightly behind the 1.38 branch in terms of fixes ...
> >
> > I'll download it in a few minutes...
> >
> >> Although I haven't been able to reproduce the pool problems you are
> >> having, I have been working on it nevertheless. When I was able to
> >> duplicate the two drive problem of not running jobs simultaneously on
> >> both drives, I took the time to review the reservation system in
> >> detail, and I found a few places where there *may* have been problems.
> >
> > I should see if the version 1.38.3beta fixes my problem on Tuesday -
> > that's still my jobs-to-different-pools-day. I'll report what happens.
> > And, of course, I've got one run of straight and usual backups before
> > that, although I don't think there will be any problems tomorrow.
>
> Unfortunately, I have to report something rather unexpected. The device
> reservation system works a little too good ;-)
>
> I started my bachup server this morning (or rather, it started itself).
> All tape drives Bacula itself loads were empty, magazines were filled
> with the correct volumes. Basically, the same situation I have 5 out of
> seven days for about a year now, with many versions of Bacula.
>
> Jobs were scheduled as usual and got triggered as usual, too.
>
> After that, not much progress since this morning, 8:20 (almost 12 hours
>
> ago):
...
> Or, in other words, the tape which is loaded is still at BOT which
> indicates that the SD didn't actually try to read it.
>
> The debug log is filled with many thousand lines mostly created by
> reserve.c in case of the mount command.
I missed this comment about the debug log. If you still have it, could you
please send it to me?
--
Best regards,
Kern
(">
/\
V_V
---
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=7637&alloc_id=16865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Cycling a USB drive - confused...
On Wednesday 14 December 2005 12:14, Greg Cope, e-DBA Team wrote:
> On Tue, 2005-12-13 at 16:51 +0100, Kern Sibbald wrote:
> > Hello,
> >
> > I've seen requests like yours before, and never really answered them
> > because I didn't really know what to do with them. However, for some
> > reason, this email peaked my interest, and I dug out an old 64MB USB
> > device that I received from Dell when I bought a computer some time ago.
> > I plugged the silly thing into my laptop and after a bit of scrounging
> > around found that all I needed to do was mount it, and presto, I had a
> > removable filesystem.
>
> I am using 250/300Gb disks in enclosures.
Yes, I imagined this was the case, and it is certainly how most people will
want to use it. In my case, I now have two small 2GB USB gadgets in hand
that I can play with.
>
> > Anyway, concerning Bacula working with these USB removable filesystems.
> > The big problem is that Bacula doesn't understand what a removable
> > filesystem is (except for DVDs) and hence expects *all* your Volumes to
> > be online all the time.
>
> Ah!
>
> > It seems to me that with a bit of tweaking, not only could Bacula be made
> > to know what is and what is not online, but it could also automatically
> > mount the Volume if needed. All I need to do is steal (well, adapt is a
> > better word) a bit of the DVD code.
>
> That would be great, but I know youre adding loads of other projects and
> the recent voting did not include this.
Well, don't hold your breath. In terms of my priorities: #1 is getting 1.38.3
working satisfactorily, #2 is integrating, testing, helping other developers,
#3 is getting the Migration code working, and #4 is projects that amuse me
such as this one. I'll probably be able to work on it a bit in January. For
the moment, #1 and #2 have me submerged.
>
> > In the mean time, dealing with these things won't be so simple for you.
> > What you need to do is write two little scripts that you run: 1. a script
> > that you run just before you remove a device, and 2. a script that you
> > run when you insert the device.
>
> Presently I can do this manually as we only swap once every few weeks.
>
> > The first script should find all the Bacula Volumes on the device (simple
> > if you only have Bacula Volumes on it) then mark them with status
> > "Disabled" in the catalog if the current catalog status is "Append".
>
> How can I do that - is there a command or bconsole argument?
I see you found it :-)
>
> > The second script would do the inverse -- for each Bacula Volume on the
> > device that is marked with status "Disabled" in the catalog, it would
> > change it to "Append".
> >
> > This would solve 90% of all the problems -- there will be additional
> > problems with Volumes that are marked Purged or Recycled. If you really
> > want to handle those correctly, in the first script, you will mark all
> > the Volumes currently on the device as "Disabled", and save their prior
> > status (and Volume name) in a file on the device. The second script
> > would then reverse the process by setting the catalog status back to what
> > it previously was.
> >
> > The above is a bit of a kludge, but I don't see any other way to do it
> > without modifying Bacula code -- perhaps some clever user knows a better
> > way. I see no reason why Bacula could not be modified to handle this
> > automatically, but that requires some code and programming time ...
>
> Thanks for the answer/thoughts.
>
> I think that a few people are trying to use USB drives, and more would
> do so if bacula supported it better.
--
Best regards,
Kern
(">
/\
V_V
---
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=7637&alloc_id=16865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Cycling a USB drive - confused...
Thanks for that the following is what I think I am after in bconsole update Volume=$VOL VolStatus=$STS I must have missed this in the docs. Greg On Wed, 2005-12-14 at 13:15 +0100, Ferdinando Pasqualetti wrote: > > I have done something silmilar with tape magazines. I use normal > backups on disks, with Full on sunday and Incr on weekdays. > > On saturday I schedule a Full copy on tapes (2 drives on a 19 slots > changer). We keep 3 magazines wich are rotated, each containing some > volumes of the 4 pools onto which we make the backups. Volumes expire > after 24 days. > > On monday we can verify if all the backups were done correctly > (eventually we have all the week to integrate failed ones). > > When we want to change the magazine we use a script (through WEBMIN) > that unmount and unload the drives, then we change the magazine and > run another script (really is the same with another argument) that > disables the volumes in the previous mag and enable the new ones. > I will attach the script, but it is not general and really not > commented. Also it is working with 1.36.3 and messages are in > Italian. > > If someone in the list is interested I could try to make it more > general and usable. > > -- > Ferdinando Pasqualetti > G.T.Dati srl > Tel. 0557310862 - 3356172731 - Fax 055720143 > > > > > [EMAIL PROTECTED] wrote on 14/12/2005 12.34.19: > > > Hi Kern, > > > > (quoting may be a bit confused; sorry) > > > > >> It seems to me that with a bit of tweaking, not only could > Bacula be > > made to > > >> know what is and what is not online, but it could also > automatically > > mount > > >> the Volume if needed. All I need to do is steal (well, adapt is > a > > better > > >> word) a bit of the DVD code. > > > > That would be very useful to me too, trying to use an Iomega Rev as > per > > earlier messages to this list. > > > > >> In the mean time, dealing with these things won't be so simple > for > > you. What > > >> you need to do is write two little scripts that you run: 1. a > script > > that you > > >> run just before you remove a device, and 2. a script that you > run > > when you > > >> insert the device. > > > > These scripts sound rather involved! Has anyone written anything > similar? > > > > Thanks again Kern and the other developers / list support crew for > an > > excellent job! > > > > -- > > Regards > > Richard (MQ) > > > > > > > > --- > > 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=7637&alloc_id=16865&op=click > > ___ > > Bacula-users mailing list > > Bacula-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/bacula-users > > --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Re: Installation Problem
Hi Scott, I've the same problem as you. Did you already solve it and can you tell me how? --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Cycling a USB drive - confused...
I have done something silmilar with tape magazines. I use normal backups on disks, with Full on sunday and Incr on weekdays. On saturday I schedule a Full copy on tapes (2 drives on a 19 slots changer). We keep 3 magazines wich are rotated, each containing some volumes of the 4 pools onto which we make the backups. Volumes expire after 24 days. On monday we can verify if all the backups were done correctly (eventually we have all the week to integrate failed ones). When we want to change the magazine we use a script (through WEBMIN) that unmount and unload the drives, then we change the magazine and run another script (really is the same with another argument) that disables the volumes in the previous mag and enable the new ones. I will attach the script, but it is not general and really not commented. Also it is working with 1.36.3 and messages are in Italian. If someone in the list is interested I could try to make it more general and usable. -- Ferdinando Pasqualetti G.T.Dati srl Tel. 0557310862 - 3356172731 - Fax 055720143 [EMAIL PROTECTED] wrote on 14/12/2005 12.34.19: > Hi Kern, > > (quoting may be a bit confused; sorry) > > >> It seems to me that with a bit of tweaking, not only could Bacula be > made to > >> know what is and what is not online, but it could also automatically > mount > >> the Volume if needed. All I need to do is steal (well, adapt is a > better > >> word) a bit of the DVD code. > > That would be very useful to me too, trying to use an Iomega Rev as per > earlier messages to this list. > > >> In the mean time, dealing with these things won't be so simple for > you. What > >> you need to do is write two little scripts that you run: 1. a script > that you > >> run just before you remove a device, and 2. a script that you run > when you > >> insert the device. > > These scripts sound rather involved! Has anyone written anything similar? > > Thanks again Kern and the other developers / list support crew for an > excellent job! > > -- > Regards > Richard (MQ) > > > > --- > 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=7637&alloc_id=16865&op=click > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > TapeSet Description: Binary data
Re: [Bacula-users] Cycling a USB drive - confused...
Hi Kern, (quoting may be a bit confused; sorry) >> It seems to me that with a bit of tweaking, not only could Bacula be made to >> know what is and what is not online, but it could also automatically mount >> the Volume if needed. All I need to do is steal (well, adapt is a better >> word) a bit of the DVD code. That would be very useful to me too, trying to use an Iomega Rev as per earlier messages to this list. >> In the mean time, dealing with these things won't be so simple for you. What >> you need to do is write two little scripts that you run: 1. a script that you >> run just before you remove a device, and 2. a script that you run when you >> insert the device. These scripts sound rather involved! Has anyone written anything similar? Thanks again Kern and the other developers / list support crew for an excellent job! -- Regards Richard (MQ) --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Cycling a USB drive - confused...
On Tue, 2005-12-13 at 16:51 +0100, Kern Sibbald wrote: > Hello, > > I've seen requests like yours before, and never really answered them because > I > didn't really know what to do with them. However, for some reason, this > email peaked my interest, and I dug out an old 64MB USB device that I > received from Dell when I bought a computer some time ago. I plugged the > silly thing into my laptop and after a bit of scrounging around found that > all I needed to do was mount it, and presto, I had a removable filesystem. I am using 250/300Gb disks in enclosures. > Anyway, concerning Bacula working with these USB removable filesystems. The > big problem is that Bacula doesn't understand what a removable filesystem is > (except for DVDs) and hence expects *all* your Volumes to be online all the > time. Ah! > It seems to me that with a bit of tweaking, not only could Bacula be made to > know what is and what is not online, but it could also automatically mount > the Volume if needed. All I need to do is steal (well, adapt is a better > word) a bit of the DVD code. That would be great, but I know youre adding loads of other projects and the recent voting did not include this. > In the mean time, dealing with these things won't be so simple for you. What > you need to do is write two little scripts that you run: 1. a script that you > run just before you remove a device, and 2. a script that you run when you > insert the device. Presently I can do this manually as we only swap once every few weeks. > The first script should find all the Bacula Volumes on the device (simple if > you only have Bacula Volumes on it) then mark them with status "Disabled" in > the catalog if the current catalog status is "Append". How can I do that - is there a command or bconsole argument? > The second script would do the inverse -- for each Bacula Volume on the > device > that is marked with status "Disabled" in the catalog, it would change it to > "Append". > > This would solve 90% of all the problems -- there will be additional problems > with Volumes that are marked Purged or Recycled. If you really want to > handle those correctly, in the first script, you will mark all the Volumes > currently on the device as "Disabled", and save their prior status (and > Volume name) in a file on the device. The second script would then reverse > the process by setting the catalog status back to what it previously was. > > The above is a bit of a kludge, but I don't see any other way to do it > without > modifying Bacula code -- perhaps some clever user knows a better way. I see > no reason why Bacula could not be modified to handle this automatically, but > that requires some code and programming time ... Thanks for the answer/thoughts. I think that a few people are trying to use USB drives, and more would do so if bacula supported it better. Greg --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Solaris 10 client change root password
Phil Stracchino wrote: > Didier Herrera wrote: >>checking host system type... sparc-sun-solaris2.8 >>passwd: Changing password for root >>New Password: >> >>It continues asking me to change the password indefinetily, what's >>wrong? I'll apreciate suggestions. > > Good question. NOTHING, I repeat NOTHING, in the Bacula build process > should be asking you to change the root password. Nor should ./configure or the build process really be run as root, I'd suggest... -- Russell Howe [EMAIL PROTECTED] --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Storing job reports as files
Hello, On Tuesday 09 August 2005 12:22, Dominic Marks wrote: Hello All, Can I make Bacula write out its report for a job to a file, or store it in a database? I'd like to import them into MySQL automatically. How I did it in the end: Direct E-Mails to an alias which delivers to a process, write a small application which receives the reports, parses them and inserts them into a table. I could have also used the file output method, but this was quicker and works over multiple hosts. I made a perl script that get the job information from the database. I have attached it if someone like to use it. report.pl Description: Perl program
Re: [Bacula-users] Catalog Backup showing error
Moin.
from the "make_catalog_backup" script:
# This script dumps your Bacula catalog in ASCII format
# It works for MySQL, SQLite, and PostgreSQL
#
# $1 is the name of the database to be backed up and the name
# of the output file (default = bacula).
# $2 is the user name with which to access the database
# (default = bacula).
# $3 is the password with which to access the database or "" if no password
# (default "")
#
Now compare that to the way you execute it.
Have a nice day,
Michael
Florian Schnabel wrote:
> Beren Gamble wrote:
>> I've got this error when bacula tries to backup the catalog:
>>
>> RunBefore: /usr/bin/mysqldump: option '-u' requires an argument
>>
>> Here's my job:
>>
>> Job {
>> Name = "BackupCatalog-Daily"
>> Type = backup
>> Level = Full
>> Client = BACKUP1-fd
>> FileSet="Catalog"
>> Schedule = "DailyCycleAfterBackup"
>> Storage = AIT50-ST1
>> Pool = AIT50-ST1-Daily
>> Messages = Standard
>> # This creates an ASCII copy of the catalog
>> RunBeforeJob = "/etc/bacula/make_catalog_backup -u bacula"
>> # This deletes the copy of the catalog
>> RunAfterJob = "/etc/bacula/delete_catalog_backup"
>> Write Bootstrap = "/var/bacula/BackupCatalog.bsr"
>> Priority = 90 # run after main backup
>> }
>>
>>
>> Any ideas?
>>
>> Also, when I get this script working, can I use it to back up other
>> mysql databases?
>>
>> Cheers,
>>
>> Beren
>>
>
> am starting with the easiest one .. you amde sure you did reload your
> configuration ?
>
> also jobs already in queue don't get affected from reload
>
> Florian
>
>
> ---
> 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=7637&alloc_id=16865&op=click
> ___
> Bacula-users mailing list
> Bacula-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bacula-users
>
--
GiS - Gesellschaft fuer integrierte Systemplanung mbH
+==+
Michael Scherer [EMAIL PROTECTED] Tel: 06201-503-74
Junkersstr.2 69469 Weinheim Fax: 06201-503-66
+==+
It is often easier to not do something dumb
than it is to do something smart.
---
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=7637&alloc_id=16865&op=click
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Re: Segmentation violation
I have tried to run several restorejobs, and every job who is beneth 516 MB i memory consumption is a sucess, every job above 516 crash whith same error message. MySQL is installed same way as when running bacula 1.36. Bacula did crash earlyer when consuming more than fysical mermory, but the box was upgraded to 2GB prior to 1.36, and the versions prior to 1.36 and 1.36 worked fine. Kern Sibbald wrote: Hello, I repeat what I said the last time: this looks like a MySQL problem. Given the new information you have presented, and the fact that you are having problems when working with almost 9 million files backed up (a lot), I would suspect that the problem is with your MySQL configuration. If you used the standard MySQL installation, it is probably not configured for such large databases. See below for more ... On Monday 12 December 2005 10:52, Roger Kvam wrote: I Tried to upgrade MySQL to 5. and recomile bacula (yes I did a make distclean), but same error, have now installed MySQL 4.1.15 again, and recompiled bacula again, deleted the database and deleted the tapes so I could start fresh. When trying to restore from a large job, bacula crashes, From the message you show below, it looks to me like MySQL crashed or at least disconnected. when trying to restore from a small job everything works fine. Does bacula has shortcomings when comes to enterprise systems? Bacula is not what I would call an enterprise system, so one could say it has that kind of shortcoming. In your case, you are dealing with a quite large backup set, and it looks like you have not adapted MySQL to handle it appropriatedly. Perhaps some of the other users on the list can help you. If I am not mistaken, the Bacula document also mentions some steps you might take for large databases, but your best bet is to look at the MySQL documentation. Now I have run several backups of several servers and everything works fine, until I wanted to do a restore from the biggest job, +---+---+---+-+-+-- --+---+ | JobId | Level | JobFiles | JobBytes| StartTime | VolumeName | StartFile | +---+---+---+-+-+-- --+---+ | 2 | F | 8,759,302 | 821,505,476,167 | 2005-12-09 16:27:02 | 09L2 | 0 | +---+---+---+-+-+-- --+---+ You have selected the following JobId: 2 Building directory tree for JobId 2 ... Query failed: SELECT MediaType FROM JobMedia,Media WHERE JobMedia.JobId=2 AND JobMedia.MediaId=Media.MediaId: ERR=Lost connection to MySQL server during query There were no files inserted into the tree, so file selection is not possible.Most likely your retention policy pruned the files Do you want to restore all the files? (yes|no): ¤¤ When trying to restore from a smaller job, everything works fine: +---+---+--++-+ +---+ | JobId | Level | JobFiles | JobBytes | StartTime | VolumeName | StartFile | +---+---+--++-+ +---+ | 3 | F | 833,832 | 52,358,007,603 | 2005-12-09 16:29:06 | 09L2 |29 | | 4 | I |1,434 |322,179,540 | 2005-12-10 00:05:01 | 01L2 | 0 | +---+---+--++-+ +---+ You have selected the following JobIds: 3,4 Building directory tree for JobId 3 ... + Building directory tree for JobId 4 ... 2 Jobs, 828,824 files inserted into the tree. Have I met any limitations in bacula? The system status was like this the moment before the crash: last pid: 34607; load averages: 1.24, 0.58, 0.25 up 3+19:19:56 08:57:34 59 processes: 6 running, 53 sleeping Mem: 572M Active, 1156M Inact, 212M Wired, 64M Cache, 112M Buf, 3008K Free Swap: 4096M Total, 500K Used, 4095M Free PID USERNAMETHR PRI NICE SIZERES STATE C TIME WCPU COMMAND 20584 mysql 9 200 57640K 33344K kserel 0 256:57 86.77% mysqld 31511 root 6 200 5608K 2304K kserel 0 144:10 0.00% bacula-sd 31532 root 6 200 518M 513M RUN0 53:32 0.00% bacula-dir Kern Sibbald wrote: This looks to me like a problem with MySQL. Did you by any chance load pre-built binaries rather than build on your system. My best guess at the moment is that Bacula is built with version X of MySQL and you are running with version Y on your machine. That means that some of the packet fields have moved or are aligned differently. Rebuiding Bacula from source directly on your machine should resolve this, if I am right ... To answer your question. Yes, of course, you can do as many backup an
[Bacula-users] Backup to file OK, but restore fails
Hi! Yesterday I upgraded my director, fd and sd from bacula 1.36.2 (that comes with Debian sarge) to 1.38.2. I have one machine running the director (with a connected tape-station) and one machine with a lot of disks and a storage daemon (this should later on be placed elsewhere, hence the separation of the daemons). The director is named merry-dir and the storage daemon is named pippin-sd. Before the upgrade I could backup for example the catalog to a file storage on the machine with just the sd. I could also restore files from the backup. After the upgrade, I could backup just as before. The job exits OK and I could see that the files that hold the backup are growing. But I cannot restore... When running a restore I could mark the files I want and the jobs is visible underneath "Running Jobs" in bconsole. The job is "waiting for Client merry-fd to connect to Storage File" and after a while the job times out. Here's the output when trying to restore a file to merry-fd (running on the same server as the director): 14-Dec 09:03 merry-dir: Start Restore Job RestoreFiles.2005-12-14_09.03.22 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Authorization key rejected by Storage daemon. Please see http://www.bacula.org/rel-manual/faq.html#AuthorizationErrors for help. 14-Dec 09:13 merry-fd: RestoreFiles.2005-12-14_09.03.22 Fatal error: Failed to authenticate Storage daemon. 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Fatal error: Socket error on Storage command: ERR=No data available 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 0 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** 14-Dec 09:13 merry-dir: RestoreFiles.2005-12-14_09.03.22 Error: Bacula 1.38.2 (20Nov05): 14-Dec-2005 09:13:30 JobId: 512 Job:RestoreFiles.2005-12-14_09.03.22 Client: merry-fd Start time: 14-Dec-2005 09:03:25 End time: 14-Dec-2005 09:13:30 Files Expected: 1 Files Restored: 0 Bytes Restored: 0 Rate: 0.0 KB/s FD Errors: 1 FD termination status: SD termination status: Waiting on FD Termination:*** Restore Error *** Note the job fails twice... Is that really correct? I could backup and restore without any problems to a local file on the server running the director (tested only once though). I could check the status of the storage daemon from the director, and also do backups so the passwords should be OK, right? I've also seen this error sometimes "14-Dec 09:43 merry-sd: Job RestoreFiles.2005-12-14_09.43.50 waiting to reserve a device." but don't really know the meaning. In a previous post i saw that Kern asked about the /lib/tls-directory when presented with this error. I do have the /lib/tls, but start all my daemons with the startscript that bacula provides. In that script the "LD_ASSUME_KERNEL=2.4.19"-variable is exported. (If running kernel 2.4! is the /lib/tls a problem with kernel 2.6 too?) I use kernel 2.4.27 on the server with the director, and 2.6.8 on the sd server. The /lib/tls is there on both the servers which are running Debian sarge. Should really "merry-sd" be involved? I've selected the storage on pippin-sd when running "restore" in bconsole. Could it be that the director/fd is trying to restore from the wrong sd? Shouldn't the job then fail with a "volume not found" error, ask me to mount a volume, or similiar? If I try to cancel a job, the director from time to time complains it cannot the job even though it's listed under running jobs. *cancel Automatically selected Job: JobId=515 Job=RestoreFiles.2005-12-14_10.18.20 Confirm cancel (yes/no): yes 3902 Job RestoreFiles.2005-12-14_10.18.20 not found. The job is marked "has been canceled" in bconsole anyway. I don't remember if I got errors like this before the upgrade. What could be wrong in my setup? I've been banging my head to the wall for quite a few hours now and I'm out of ideas. Best regards, Jonas Mixter --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/list
[Bacula-users] Re: Trouble using SSH tunnel with Bacula
Hello Joshua, On Tue, 13 Dec 2005 08:56:40 -0900 Joshua Kugler <[EMAIL PROTECTED]> wrote: > Well, what do you know. :) That worked. It would appear that the > ssh-tunnel example script needs some work (I also had to change some > other things to get it to find the PID correctly). I'll write up a > little document on how I got this up and running, and include my > version of the ssh-tunnel script. I am fine with this! Please feel free to add a modified version of the ssh-tunnel.sh script to the sources. Thanks for working on this Stephan -- Stephan Holl Check headers for GnuPG Key! http://www.gdf-hannover.de signature.asc Description: PGP signature
Re: [Bacula-users] FileDaemon vs manual transfer
-- Ferdinando Pasqualetti G.T.Dati srl Tel. 0557310862 - 3356172731 - Fax 055720143 Florian Daniel Otel <[EMAIL PROTECTED]> wrote on 14/12/2005 09.28.11: > On 12/13/05, Ferdinando Pasqualetti <[EMAIL PROTECTED]> wrote: > > > > I was giving that suggestion because, as far as I know bacula is naot > > capable of resuming a job aborted by a broken connection. > > So, if I read you correctly, a temporary drop in the connection btw. > the FileDaemon and StorageDaemon will cause the job to be aborted. Any > workarounds (in bacula) ? AFAIK the only way you can try (to avoid the connection breaking, not to resume the job) is to setup a "Hearthbeat Interval = someseconds", if this can help. > > > Florian > --- 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=7637&alloc_id=16865&op=click ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Re: Bacula BETA 1.38.3
Hello, some more results. I was curious, so I found the opportunity to verify at least part of todays regular backup run. Arno Lehmann schrieb: Kern - Kern Sibbald schrieb: Hmmm. I think I am going to pack it in and take a year's vacation ... don't. I was a little stupid when I tried to analyze the problem. Or rather, I didn't really analyze. What I did this evening: Shutdown bacula. Use mtx to load a tape into ONE of my TWO autoloaders (might be important later). Start Bacula again. Start a job going to autoloader ONE. Voilà. New results now: When Bacula encounters an empty autoloader drive at startup, I can't get it loading and using tapes. When the drive is loaded when Bacula (SD, probably) starts, everything's ok. What I will / might / should test: - What happens with the empty autoloader TWO? I suspect it will remain blocked. - What happens if I umount ONE and later mount it again? I had to change tapes. The normal umount - change magazines - update slots scan - mount worked. Jobs that go to the loaded volumes run, others wait. No jobs failed until now, but up till now my original problem situation didn't happen. Tomorrow, I've got a job scheduled that should go to TWO - I'll report. This job is still waiting for Maximum Concurrent Jobs on the client... Arno Arno On Monday 12 December 2005 20:21, Arno Lehmann wrote: Hi, Arno Lehmann schrieb: Hi, Kern Sibbald schrieb: ... Anyway, I would really appreciate it if you could try the 1.38.3 beta. Please use the released tar file rather than the CVS -- the HEAD CVS is now slightly behind the 1.38 branch in terms of fixes ... I'll download it in a few minutes... Although I haven't been able to reproduce the pool problems you are having, I have been working on it nevertheless. When I was able to duplicate the two drive problem of not running jobs simultaneously on both drives, I took the time to review the reservation system in detail, and I found a few places where there *may* have been problems. I should see if the version 1.38.3beta fixes my problem on Tuesday - that's still my jobs-to-different-pools-day. I'll report what happens. And, of course, I've got one run of straight and usual backups before that, although I don't think there will be any problems tomorrow. Unfortunately, I have to report something rather unexpected. The device reservation system works a little too good ;-) I started my bachup server this morning (or rather, it started itself). All tape drives Bacula itself loads were empty, magazines were filled with the correct volumes. Basically, the same situation I have 5 out of seven days for about a year now, with many versions of Bacula. Jobs were scheduled as usual and got triggered as usual, too. After that, not much progress since this morning, 8:20 (almost 12 hours ago): 12-Dec 11:21 goblin-sd: Job DracheStd.2005-12-12_08.20.00 waiting to reserve a device. 12-Dec 11:26 goblin-sd: Job Goblin.2005-12-12_08.25.00 waiting to reserve a device. 12-Dec 11:26 goblin-sd: Job Ork.2005-12-12_08.25.03 waiting to reserve a device. 12-Dec 11:26 goblin-sd: Job ElfSys.2005-12-12_08.25.01 waiting to reserve a device. 12-Dec 11:26 goblin-sd: Job ElfHome.2005-12-12_08.25.02 waiting to reserve a device. 12-Dec 15:21 goblin-sd: Job DracheStd.2005-12-12_08.20.00 waiting to reserve a device. 12-Dec 15:26 goblin-sd: Job Goblin.2005-12-12_08.25.00 waiting to reserve a device. 12-Dec 15:26 goblin-sd: Job Ork.2005-12-12_08.25.03 waiting to reserve a device. 12-Dec 15:26 goblin-sd: Job ElfSys.2005-12-12_08.25.01 waiting to reserve a device. 12-Dec 15:26 goblin-sd: Job ElfHome.2005-12-12_08.25.02 waiting to reserve a device. these are the latest messages. Or, in other words, since I upgraded to 1.38.3 beta, nothing has been backed up at all. The current director status is #sta dir goblin-dir Version: 1.38.3 (09 December 2005) i586-pc-linux-gnu suse 8.1 Daemon started 12-Dec-05 07:32, 1 Job run since started. Scheduled Jobs: Level Type Pri Scheduled Name Volume = == IncrementalBackup10 13-Dec-05 08:20 DracheStdDLT-IV-0063 IncrementalBackup10 13-Dec-05 08:20BeowulfStd DLT-IV-0063 Differential Backup10 13-Dec-05 08:25 Goblin DLT-IV-0063 Differential Backup10 13-Dec-05 08:25ElfSys DLT-IV-0063 Differential Backup10 13-Dec-05 08:25ElfHomeDLT-IV-0063 Differential Backup 10 13-Dec-05 08:25Ork*unknown* Differential Backup30 13-Dec-05 08:25GoblinDB DLT-IV-0063 Differential Backup10 13-Dec-05 08:30BackupMail QIC-525-0046 Full Backup 100 13-Dec-05 08:33 BackupCatalog DAT-120-0006 Admin999 13-Dec-05 09:00 Shutdown Running Jobs: JobId Level Name
