Re: [Bacula-users] Job is not progressing - tape mounted
In addition, the tape in the drive is: $ echo 'list media' | bconsole | grep DLT7000-017 | 60 | DLT7000-017 | Recycle | 1 | 1 | 0 | 63,072,000 | 1 |0 | 1 | DLT | 2008-02-20 16:34:58 | -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Job is not progressing - tape mounted
I am not sure why this job is not progressing. I have tried umount/mount and things don't progress. No messages from the console. I am not running concurrent jobs. Recent changes (week or so ago) include duplicate job control. Bacula 3.0.3 on FreeBSD 8.x client running now is also 3.0.3 on FreeBSD 8.x after a umount: Device status: Autochanger "DEC TL800" with devices: DEC TZ89 Device "DEC TZ89" is not open or does not exist. Device "FileStorage" (/bacula-volumes/bacula/volumes) is not open. Device "DLT" (/dev/nsa0) is not open. Device is BLOCKED. User unmounted during wait for media/mount. After a mount: *mount storage=DLT 3001 Device "DLT" (/dev/nsa0) is mounted with Volume "DLT7000-017" *status storage=DLT [snip] Device status: Autochanger "DEC TL800" with devices: DEC TZ89 Device "DEC TZ89" is not open or does not exist. Device "FileStorage" (/bacula-volumes/bacula/volumes) is not open. Device "DLT" (/dev/nsa0) is mounted with: Volume: DLT7000-017 Pool:Default Media type: DLT Device is BLOCKED waiting to create a volume for: Pool:Default Media type: DLT Total Bytes Read=64,512 Blocks Read=1 Bytes/block=64,512 Positioned at File=0 Block=0 Full status outputs follow: Director status: $ bconsole Connecting to Director bacula.exampleorg:9101 1000 OK: bacula-dir Version: 3.0.3 (18 October 2009) Enter a period to cancel a command. *status dir bacula-dir Version: 3.0.3 (18 October 2009) i386-portbld-freebsd8.0 freebsd 8.0-STABLE Daemon started 07-Dec-09 12:22, 31 Jobs run since started. Heap: heap=0 smbytes=501,531 max_bytes=502,084 bufs=1,795 max_bufs=2,192 Scheduled Jobs: Level Type Pri Scheduled Name Volume === IncrementalBackup20 10-Dec-09 04:00nyi maildir DLT7000-016 IncrementalBackup10 10-Dec-09 05:55wocker DLT7000-016 IncrementalBackup10 10-Dec-09 05:55polo DLT7000-016 IncrementalBackup10 10-Dec-09 05:55polo basic DLT7000-016 IncrementalBackup10 10-Dec-09 05:55dbclone home DLT7000-016 IncrementalBackup10 10-Dec-09 05:55ngaio basic DLT7000-016 IncrementalBackup10 10-Dec-09 05:55dbclone basic DLT7000-016 IncrementalBackup10 10-Dec-09 05:55bast home DLT7000-016 IncrementalBackup10 10-Dec-09 05:55bast basic DLT7000-016 IncrementalBackup10 10-Dec-09 05:55ngaio databases DLT7000-016 IncrementalBackup10 10-Dec-09 05:55ngaio DLT7000-016 IncrementalBackup20 10-Dec-09 05:55nyi DLT7000-016 IncrementalBackup20 10-Dec-09 08:00nyi maildir DLT7000-016 IncrementalBackup20 10-Dec-09 08:15supernews DLT7000-016 IncrementalBackup20 10-Dec-09 08:15latens home DLT7000-016 IncrementalBackup20 10-Dec-09 08:15latens basic DLT7000-016 Full Backup 100 10-Dec-09 08:15BackupCatalog DLT7000-016 Differential Backup20 10-Dec-09 12:01nyi maildir DLT7000-016 Full Backup20 10-Dec-09 12:01nyi maildir tarball DLT7000-016 IncrementalBackup20 10-Dec-09 16:00nyi maildir DLT7000-016 IncrementalBackup20 10-Dec-09 20:00nyi maildir DLT7000-016 IncrementalBackup20 11-Dec-09 00:00nyi maildir DLT7000-016 Running Jobs: Console connected at 10-Dec-09 03:28 JobId Level Name Status == 31443 Increme nyi_maildir.2009-12-09_00.00.00_30 is waiting for higher priority jobs to finish 31446 Increme bast_basic.2009-12-09_05.55.01_33 is running 31447 Increme bast_home.2009-12-09_05.55.01_34 is waiting on max Storage jobs 31448 Increme dbclone_basic.2009-12-09_05.55.01_35 is waiting on max Storage jobs 31449 Increme dbclone_home.2009-12-09_05.55.01_36 is waiting on max Storage jobs 31450 Increme polo_basic.2009-12-09_05.55.01_37 is waiting on max Storage jobs 31451 Increme polo.2009-12-09_05.55.01_38 is waiting on max Storage jobs 31452 Increme ngaio_basic.2009-12-09_05.55.02_39 is waiting on max Storage jobs 31453 Increme ngaio_databases.2009-12-09_05.55.02_40 is waiting on max Storage jobs 31454 Increme ngaio.2009-12-09_05.55.02_41 is waiting on max Storage jobs 31455 Increme nyi.2009-12-09_05.55.02_42 is waiting execution 31457 Increme supernews.2009-12-09_08.15.00_44 is waiting execution 31458 Increme latens_basic.2009-12-09_08.15.00_45 is waiting execution 31459 Increme latens_home.2009-12-09_08.15.00_46 is waiting execution 31460 FullBackupCatalog.2009-12-09_08.15.00_47 is waiting execution 31462 Fullnyi_maildir_tarball.2009-12-09_12.01.00_49 is waiting execution Terminated Jobs: JobId LevelFiles Bytes Status FinishedName
[Bacula-users] backup error - Network send error to SD
Hi everyone. I have a new problem with a particular backup in my environment. I'm using bacula 3.0.2 across all components. This has been running fine for some time however this particular laptop used to be running linux(recently it was switched over to XP). This laptop is recently routinely getting the following backup error: 09-Dec 14:04 pendual-dir JobId 9: Start Backup JobId 9, +Job=ncs6320-backup.2009-12-09_14.04.49_03 09-Dec 14:04 pendual-dir JobId 9: Using Device "Drive-1" 09-Dec 14:04 pendual-sd JobId 9: Volume "125m_1_1" previously written, moving to +end of data. 09-Dec 14:06 pendual-sd JobId 9: Ready to append to end of Volume "125m_1_1" at +file=16. 09-Dec 14:06 pendual-sd JobId 9: Spooling data ... 09-Dec 14:06 ncs6320-fd JobId 9: Generate VSS snapshots. Driver="VSS WinXP", +Drive(s)="C" 09-Dec 14:17 ncs6320-fd JobId 9: Fatal error: +/home/kern/bacula/k/src/filed/backup.c:964 Network send error to SD. +ERR=Input/output error 09-Dec 14:19 ncs6320-fd JobId 9: VSS Writer (BackupComplete): "Microsoft Writer +(Bootable State)", State: 0x1 (VSS_WS_STABLE) 09-Dec 14:19 ncs6320-fd JobId 9: VSS Writer (BackupComplete): "MSDEWriter", +State: 0x1 (VSS_WS_STABLE) 09-Dec 14:19 ncs6320-fd JobId 9: VSS Writer (BackupComplete): "WMI Writer", +State: 0x1 (VSS_WS_STABLE) 09-Dec 14:19 ncs6320-fd JobId 9: VSS Writer (BackupComplete): "Microsoft Writer +(Service State)", State: 0x1 (VSS_WS_STABLE) 09-Dec 14:19 pendual-sd JobId 9: Job ncs6320-backup.2009-12-09_14.04.49_03 +marked to be canceled. 09-Dec 14:19 pendual-dir JobId 9: Error: Bacula pendual-dir 3.0.2 (18Jul09): +09-Dec-2009 14:19:10 estimate jobs run fine. Short test backup jobs run fine(like just backing up a single file). The point in the backup that it fails at seems to be random. 8 Full 1,65650.53 M Error09-Dec-09 13:52 ncs6320-backup 9 Full 20,8311.030 G Error09-Dec-09 14:27 ncs6320-backup 11 Full 2,29460.65 M Error09-Dec-09 14:28 ncs6320-backup 12 Full 9,080357.6 M Error09-Dec-09 14:52 ncs6320-backup I have turned on debug in the SD and the FD and see this: SD Date label written: 07-Dec-2009 14:13^M pendual-sd: label.c:236-0 Call reserve_volume=125m_1_1^M pendual-sd: bnet.c:670-0 who=client host=192.168.25.100 port=36643^M pendual-sd: cram-md5.c:73-0 send: auth cram-md5 <223312495.1260387...@pendual-sd > ssl=0^M pendual-sd: cram-md5.c:133-0 cram-get received: auth cram-md5 <788566849.1260387 9...@pendual-dir> ssl=0^M pendual-sd: cram-md5.c:152-0 sending resp to challenge: en+f7Vhrsw8IjS/4V3/A4B^M pendual-sd: dircmd.c:211-0 Message channel init completed.^M for the FD I see this: ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:123-12 Enter convert _unix_to_win32_path ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:174-12 path=C:\Docum ents and Settings\mrc\My Documents\mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:183-12 Leave cvt_u_t o_win32_path path=\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy1\Documents and Settings\mrc\My Documents\mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:123-12 Enter convert _unix_to_win32_path ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:174-12 path=C:\Docum ents and Settings\mrc\My Documents\mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:183-12 Leave cvt_u_t o_win32_path path=\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy1\Documents and Settings\mrc\My Documents\mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/findlib/bfile.c:480-12 === NO plugin ncs6320-fd: /home/kern/bacula/k/src/findlib/bfile.c:572-12 Read CreateFileW=\\?\ GLOBALROOT\Device\HarddiskVolumeShadowCopy1\Documents and Settings\mrc\My Docume nts\mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=0 stop=0 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:91-0 Got BNET_SIG 0 from S D ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:96-0 wait_intr=1 stop=1 ncs6320-fd: /home/kern/bacula/k/src/win32/compat/compat.cpp:924-12 sizino=8 ino= 0 file=C:/Documents and Settings/mrc/My Documents/mwm4-49.mp3 ncs6320-fd: /home/kern/bacula/k/src/filed/heartbeat.c:142-12 Send kill to heartb eat id ncs6320-fd: /home/kern/bacula/k/src/filed/backup.c:189-12 end blast_data ok=0 ncs6320-fd: /home/kern/bacula/k/src/filed/job.c:272-12 Quit command loop. Cancel ed=1 ncs6320-fd: /home/kern/bacula/k/src/filed/job.c:375-12 Calling term_find_files ncs6320-fd: /home/kern/bacula/k/src/filed/job.c:378-12 Done with term_find_files ncs6320-fd: /home/kern/bacula/k/src/win32
Re: [Bacula-users] Windows 2008 - Can't see restore files
I'm stupid...
I have seen the "C:/", but as I'm used too see the root ("/") on linux I
thought that it was the root for windows, but in the backup it is a normal
directory on the "/" !!
So I have to enter in the "C:/" and after I can see all my backups
Sorry and thanks!!
2009/12/9 Timo Neuvonen
>
> "Carlo Filippetto" kirjoitti viestissä
> news:8791c1920912090626o4735d964v7eb1031c885a0...@mail.gmail.com...
> > Hi,
> > I'm testing to backup two Windows 2008, the first 32 bit, the second
> 64bit
> >
> > In both case the backup jobs go well, but if I tried to 'restore' the
> > files I can't see them as I post here:
> >
> >
> > -
> > Automatically selected FileSet: Windows-bacula
> >
> +---+---+--++-++
> > | JobId | Level | JobFiles | JobBytes | StartTime |
> VolumeName
> > |
> >
> +---+---+--++-++
> > |21 | F |3 | 14,150,553 | 2009-12-09 15:04:33 |
> > VOL-daily-0004 |
> >
> +---+---+--++-++
> > You have selected the following JobId: 21
> >
> > Building directory tree for JobId(s) 21 ...
> > 2 files inserted into the tree.
> >
> > You are now entering file selection mode where you add (mark) and
> > remove (unmark) files to be restored. No files are initially added,
> unless
> > you used the "all" keyword on the command line.
> > Enter "done" to leave this mode.
> >
> > cwd is: /
> > $ ls
> > C:/
> > $ dir
> > -- 0 root root 0 1970-01-01 01:00:00 C:/
> > $ mark
> > No files marked.
> > $ mark *
> > 4 files marked.
> > $ exit
> > --
> >
> > As you can read if I ask to list the directory I can't see nothing, and
> > show the date of 1970-01-01,
> > If I mark '*' bacula marks 4 files.. and the restore is good!
> > WHY??
> > What can I do?
>
> cwd is: /
> $ ls
> D:/
> $ cd D:/
> cwd is: D:/
> $ ls
> Download/
> $ mark Download
> 35 files marked.
> $
>
> Shortly: add the file/directory name after "mark". Or add the wildcard *,
> to
> mark all. Or "cd" to a lower level directory first and...
>
> Then finish with "done", not "exit" (maybe it works too, I don't know)
>
> --
> TiN
>
>
>
>
> --
> Return on Information:
> Google Enterprise Search pays you back
> Get the facts.
> http://p.sf.net/sfu/google-dev2dev
> ___
> Bacula-users mailing list
> Bacula-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bacula-users
>
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Rate limiting in bacula?
> > Anyone can recommend a application for Microsoft Windows to limit the > bandwidth? Looking for something that is as invisible as possible for > the user and only limit the bacula-fd process or port. Freeware or open > source of course.. > Ideally, from a network architecture point of view, that is where you would do the rate limiting, but it works almost as well when done on the receiving end - you just drop the occasional packet to make the sender think that there is congestion and it lowers the sending speed accordingly. With appropriate QoS rules on the border router of the site where the SD is, you should be able to get the backups using (say) 80% of the idle bandwidth without noticing a thing, especially on a 20mbit connection where the per-packet times are really small. If your border router isn't up to the task or if you really do want to clamp the SD traffic to a fixed rate, you should still be able to just do it on the SD too. James -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
On Wed, 2009-12-09 at 14:49 -0500, Shawn wrote:
> On Wed, 2009-12-09 at 14:32 -0500, John Drescher wrote:
>
> > 2009/12/9 Shawn
> > >
> > > On Wed, 2009-12-09 at 11:34 -0600, lars_nor...@mcafee.com wrote:
> > >
> > > > -Original Message-
> > > > From: Shawn [mailto:sh...@artemide.us]
> > > > Sent: Wednesday, December 09, 2009 12:13 PM
> > > > To: bacula-users@lists.sourceforge.net
> > > > Subject: Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole
> > > > is hung up.
> > > >
> > > > Hello again, folk,
> > > >
> > > > Checked out there and there's really only a few negligible results
> > > > on this problem.
> > > >
> > > > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the
> > > > bconsole program just won't let me connect.
> > > >
> > > > I've checked both .conf files, passwords/addresses/etc, tried
> > > > changing address on both to name, IP, etc. but it won't let me connect.
> > > >
> > > > Output of:
> > > >
> > > > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt
> > > >
> > > > Connecting to Director 10.6.4.238:9101
> > > > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip
> > > > address hidden):9101] All host[ipv4:(ip address hidden):9101]
> > > > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon
> > > > host=(ip address hidden) port=9101
> > > > (Gets hung here for about 300 seconds or so)
> > > > Director authorization problem.
> > > > Most likely the passwords do not agree.
> > > > If you are using TLS, there may have been a certificate validation
> > > > error during the TLS handshake.
> > > > Please see http://www.bacula.org/en/rel-
> > > > manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for
> > > > help.
> > > >
> > > >
> > > > All I see that's noticeable, while debugging bacula-dir, is that
> > > > the client tries to connect, and that's about it:
> > > >
> > > > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client
> > > > host=(ip address hidden) port=36131
> > > >
> > > > I'm actually not using TLS or anything, has anyone encountered this
> > > > oddity before?
> > > >
> > > > Thanks in advance,
> > >
> > >
> > > Have you done a packet trace on the bacula server? May be (guessing here)
> > > as a part of the upgrade that host-based firewalling got enabled?
> > >
> > > --
> > > Return on Information:
> > > Google Enterprise Search pays you back
> > > Get the facts.
> > > http://p.sf.net/sfu/google-dev2dev
> > > ___
> > > Bacula-users mailing list
> > > Bacula-users@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/bacula-users
> > >
> > > Lars,
> > >
> > > Well, truth be told, I was doing this locally on the server. For
> > > sake of experimentation, though, I ensured bconsole.conf was proper on my
> > > desktop, and ran bconsole while doing a trace.
> > >
> > > All that happens, is five packets:
> > >
> > > 47902 > bacula-dir [SYN]
> > >
> > > Server acknowledges:
> > >
> > > bacula-dir > 47902 [SYN, ACK]
> > >
> > > 47902 > bacula-dir [ACK]
> > >
> > > Client responds with call to director:
> > >
> > > 47902 > bacula-dir [PSH, ACK]
> > >
> > > 00 24 81 f8 08 08 00 1d 92 b6 98 70 08 00 45 00 .$.. ...p..E.
> > > 0010 00 52 c7 a5 40 00 40 06 55 18 0a 06 04 ef 0a 06 @.@. U...
> > > 0020 04 ee bb 1e 23 8d 86 8e 1c 49 43 cc 0b ef 80 18 #... .IC.
> > > 0030 00 2e 4b 5d 00 00 01 01 08 0a 00 76 db eb 0f 6e ..K] ...v...n
> > > 0040 7b cc 00 00 00 1a 48 65 6c 6c 6f 20 2a 55 73 65 {.He llo *Use
> > > 0050 72 41 67 65 6e 74 2a 20 63 61 6c 6c 69 6e 67 0a rAgent* calling.
> > >
> > > Server Acknowledges...then, nothing happens at all after this:
> > >
> > > bacula-dir > 47902 [ACK]
> > >
> > > 00 1d 92 b6 98 70 00 24 81 f8 08 08 08 00 45 00 .p.$ ..E.
> > > 0010 00 34 70 23 40 00 40 06 ac b8 0a 06 04 ee 0a 06 @.@.
> > > 0020 04 ef 23 8d bb 1e 43 cc 0b ef 86 8e 1c 67 80 10 ..#...C. .g..
> > > 0030 00 2e 1f ae 00 00 01 01 08 0a 0f 6e 7b cc 00 76 ...n{..v
> > > 0040 db eb ..
> > >
> > > So what's puzzling here, is that the server is listening and even
> > > responding, yet it seems like it's hung on the response for some reason.
> > > That, or perhaps there's some communication problem in the package of
> > > bacula-console I'm working with, both in ressu's 3.0.2 and the original
> > > 2.4.4, with this director.
> > >
> >
> > run the director directly from the shell and examine the messages printed.
> >
> > # bacula-dir -d 200 -c /etc/bacula-dir.conf
> >
> > John
>
>
> John,
>
> Here's the debug level 200...again, I really don't see what's
> holding things up here:
>
> bacula-dir: sql_create.c:341-0 In create mediatype
> bacula-dir: sql_cr
Re: [Bacula-users] Network transfer Speed
Sorry, Left out the tape drive. It is a Dell TL2000 connected via fiber channel. Using 10 LTO-4 tapes. The tape drive is connected to a QLogic 4Gb Fibre Channel card. -H -Original Message- From: Gavin McCullagh [mailto:gavin.mccull...@gcd.ie] Sent: Wednesday, December 09, 2009 11:38 AM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed On Wed, 09 Dec 2009, Hayden Katzenellenbogen wrote: > I started an 800G restore. Which is half of my 1.7TB backup. > > I ran iostat while it was restoring and was getting between 50 and 140 > Meg/s. I would say the average is around 85Meg/s. Is that 85MByte/sec across the network or from local backup to local disk? Do you mind if I ask what media you're using at each end (tapes or disks, sata/scsi/lto-X/). That sounds pretty quick to me. Gavin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Gavin, Here is a quick summary. This is a single machine it is running file, storage and director daemon. The postgres database is on a separate machine. All services are communicating via IP not the loopback. There does not seem to be any difference in speed if I use loopback or IP. This is a dual quad core xeon with 4Gig of mem We have 8 SATA drives connected to a 3ware 9690SA-8I raid controller in a Raid-6 configuration. Card has 512M RAM and a battery backup unit. The issue I am running into is during a full backup I am getting around 9.5Meg/s vs. the 85Meg/s on a restore. You would think the raid array would be able to read faster than it can write. -H -Original Message- From: Gavin McCullagh [mailto:gavin.mccull...@gcd.ie] Sent: Wednesday, December 09, 2009 11:38 AM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed On Wed, 09 Dec 2009, Hayden Katzenellenbogen wrote: > I started an 800G restore. Which is half of my 1.7TB backup. > > I ran iostat while it was restoring and was getting between 50 and 140 > Meg/s. I would say the average is around 85Meg/s. Is that 85MByte/sec across the network or from local backup to local disk? Do you mind if I ask what media you're using at each end (tapes or disks, sata/scsi/lto-X/). That sounds pretty quick to me. Gavin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
On Wed, 2009-12-09 at 14:32 -0500, John Drescher wrote:
> 2009/12/9 Shawn
> >
> > On Wed, 2009-12-09 at 11:34 -0600, lars_nor...@mcafee.com wrote:
> >
> > > -Original Message-
> > > From: Shawn [mailto:sh...@artemide.us]
> > > Sent: Wednesday, December 09, 2009 12:13 PM
> > > To: bacula-users@lists.sourceforge.net
> > > Subject: Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole
> > > is hung up.
> > >
> > > Hello again, folk,
> > >
> > > Checked out there and there's really only a few negligible results
> > > on this problem.
> > >
> > > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the
> > > bconsole program just won't let me connect.
> > >
> > > I've checked both .conf files, passwords/addresses/etc, tried
> > > changing address on both to name, IP, etc. but it won't let me connect.
> > >
> > > Output of:
> > >
> > > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt
> > >
> > > Connecting to Director 10.6.4.238:9101
> > > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip
> > > address hidden):9101] All host[ipv4:(ip address hidden):9101]
> > > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon
> > > host=(ip address hidden) port=9101
> > > (Gets hung here for about 300 seconds or so)
> > > Director authorization problem.
> > > Most likely the passwords do not agree.
> > > If you are using TLS, there may have been a certificate validation
> > > error during the TLS handshake.
> > > Please see http://www.bacula.org/en/rel-
> > > manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for
> > > help.
> > >
> > >
> > > All I see that's noticeable, while debugging bacula-dir, is that
> > > the client tries to connect, and that's about it:
> > >
> > > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client
> > > host=(ip address hidden) port=36131
> > >
> > > I'm actually not using TLS or anything, has anyone encountered this
> > > oddity before?
> > >
> > > Thanks in advance,
> >
> >
> > Have you done a packet trace on the bacula server? May be (guessing here)
> > as a part of the upgrade that host-based firewalling got enabled?
> >
> > --
> > Return on Information:
> > Google Enterprise Search pays you back
> > Get the facts.
> > http://p.sf.net/sfu/google-dev2dev
> > ___
> > Bacula-users mailing list
> > Bacula-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bacula-users
> >
> > Lars,
> >
> > Well, truth be told, I was doing this locally on the server. For sake
> > of experimentation, though, I ensured bconsole.conf was proper on my
> > desktop, and ran bconsole while doing a trace.
> >
> > All that happens, is five packets:
> >
> > 47902 > bacula-dir [SYN]
> >
> > Server acknowledges:
> >
> > bacula-dir > 47902 [SYN, ACK]
> >
> > 47902 > bacula-dir [ACK]
> >
> > Client responds with call to director:
> >
> > 47902 > bacula-dir [PSH, ACK]
> >
> > 00 24 81 f8 08 08 00 1d 92 b6 98 70 08 00 45 00 .$.. ...p..E.
> > 0010 00 52 c7 a5 40 00 40 06 55 18 0a 06 04 ef 0a 06 @.@. U...
> > 0020 04 ee bb 1e 23 8d 86 8e 1c 49 43 cc 0b ef 80 18 #... .IC.
> > 0030 00 2e 4b 5d 00 00 01 01 08 0a 00 76 db eb 0f 6e ..K] ...v...n
> > 0040 7b cc 00 00 00 1a 48 65 6c 6c 6f 20 2a 55 73 65 {.He llo *Use
> > 0050 72 41 67 65 6e 74 2a 20 63 61 6c 6c 69 6e 67 0a rAgent* calling.
> >
> > Server Acknowledges...then, nothing happens at all after this:
> >
> > bacula-dir > 47902 [ACK]
> >
> > 00 1d 92 b6 98 70 00 24 81 f8 08 08 08 00 45 00 .p.$ ..E.
> > 0010 00 34 70 23 40 00 40 06 ac b8 0a 06 04 ee 0a 06 @.@.
> > 0020 04 ef 23 8d bb 1e 43 cc 0b ef 86 8e 1c 67 80 10 ..#...C. .g..
> > 0030 00 2e 1f ae 00 00 01 01 08 0a 0f 6e 7b cc 00 76 ...n{..v
> > 0040 db eb ..
> >
> > So what's puzzling here, is that the server is listening and even
> > responding, yet it seems like it's hung on the response for some reason.
> > That, or perhaps there's some communication problem in the package of
> > bacula-console I'm working with, both in ressu's 3.0.2 and the original
> > 2.4.4, with this director.
> >
>
> run the director directly from the shell and examine the messages printed.
>
> # bacula-dir -d 200 -c /etc/bacula-dir.conf
>
> John
John,
Here's the debug level 200...again, I really don't see what's
holding things up here:
bacula-dir: sql_create.c:341-0 In create mediatype
bacula-dir: sql_create.c:344-0 selectmediatype: SELECT
MediaTypeId,MediaType FROM MediaType WHERE MediaType='File'
bacula-dir: mysql.c:236-0 closedb ref=0 connected=1 db=a89740
bacula-dir: mysql.c:240-0 close db=a89740
skydro-dir: pythonlib.c:104-0 No script dir. prog=DirStartUp
skydro-dir: dird.c:317-0 Start UA server
Re: [Bacula-users] Network transfer Speed
On Wed, 09 Dec 2009, Hayden Katzenellenbogen wrote: > I started an 800G restore. Which is half of my 1.7TB backup. > > I ran iostat while it was restoring and was getting between 50 and 140 > Meg/s. I would say the average is around 85Meg/s. Is that 85MByte/sec across the network or from local backup to local disk? Do you mind if I ask what media you're using at each end (tapes or disks, sata/scsi/lto-X/). That sounds pretty quick to me. Gavin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
2009/12/9 Shawn
>
> On Wed, 2009-12-09 at 11:34 -0600, lars_nor...@mcafee.com wrote:
>
> > -Original Message-
> > From: Shawn [mailto:sh...@artemide.us]
> > Sent: Wednesday, December 09, 2009 12:13 PM
> > To: bacula-users@lists.sourceforge.net
> > Subject: Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole
> > is hung up.
> >
> > Hello again, folk,
> >
> > Checked out there and there's really only a few negligible results
> > on this problem.
> >
> > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the
> > bconsole program just won't let me connect.
> >
> > I've checked both .conf files, passwords/addresses/etc, tried
> > changing address on both to name, IP, etc. but it won't let me connect.
> >
> > Output of:
> >
> > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt
> >
> > Connecting to Director 10.6.4.238:9101
> > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip
> > address hidden):9101] All host[ipv4:(ip address hidden):9101]
> > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon
> > host=(ip address hidden) port=9101
> > (Gets hung here for about 300 seconds or so)
> > Director authorization problem.
> > Most likely the passwords do not agree.
> > If you are using TLS, there may have been a certificate validation
> > error during the TLS handshake.
> > Please see http://www.bacula.org/en/rel-
> > manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for
> > help.
> >
> >
> > All I see that's noticeable, while debugging bacula-dir, is that
> > the client tries to connect, and that's about it:
> >
> > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client
> > host=(ip address hidden) port=36131
> >
> > I'm actually not using TLS or anything, has anyone encountered this
> > oddity before?
> >
> > Thanks in advance,
>
>
> Have you done a packet trace on the bacula server? May be (guessing here) as
> a part of the upgrade that host-based firewalling got enabled?
>
> --
> Return on Information:
> Google Enterprise Search pays you back
> Get the facts.
> http://p.sf.net/sfu/google-dev2dev
> ___
> Bacula-users mailing list
> Bacula-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bacula-users
>
> Lars,
>
> Well, truth be told, I was doing this locally on the server. For sake of
> experimentation, though, I ensured bconsole.conf was proper on my desktop,
> and ran bconsole while doing a trace.
>
> All that happens, is five packets:
>
> 47902 > bacula-dir [SYN]
>
> Server acknowledges:
>
> bacula-dir > 47902 [SYN, ACK]
>
> 47902 > bacula-dir [ACK]
>
> Client responds with call to director:
>
> 47902 > bacula-dir [PSH, ACK]
>
> 00 24 81 f8 08 08 00 1d 92 b6 98 70 08 00 45 00 .$.. ...p..E.
> 0010 00 52 c7 a5 40 00 40 06 55 18 0a 06 04 ef 0a 06 @.@. U...
> 0020 04 ee bb 1e 23 8d 86 8e 1c 49 43 cc 0b ef 80 18 #... .IC.
> 0030 00 2e 4b 5d 00 00 01 01 08 0a 00 76 db eb 0f 6e ..K] ...v...n
> 0040 7b cc 00 00 00 1a 48 65 6c 6c 6f 20 2a 55 73 65 {.He llo *Use
> 0050 72 41 67 65 6e 74 2a 20 63 61 6c 6c 69 6e 67 0a rAgent* calling.
>
> Server Acknowledges...then, nothing happens at all after this:
>
> bacula-dir > 47902 [ACK]
>
> 00 1d 92 b6 98 70 00 24 81 f8 08 08 08 00 45 00 .p.$ ..E.
> 0010 00 34 70 23 40 00 40 06 ac b8 0a 06 04 ee 0a 06 @.@.
> 0020 04 ef 23 8d bb 1e 43 cc 0b ef 86 8e 1c 67 80 10 ..#...C. .g..
> 0030 00 2e 1f ae 00 00 01 01 08 0a 0f 6e 7b cc 00 76 ...n{..v
> 0040 db eb ..
>
> So what's puzzling here, is that the server is listening and even
> responding, yet it seems like it's hung on the response for some reason.
> That, or perhaps there's some communication problem in the package of
> bacula-console I'm working with, both in ressu's 3.0.2 and the original
> 2.4.4, with this director.
>
run the director directly from the shell and examine the messages printed.
# bacula-dir -d 200 -c /etc/bacula-dir.conf
John
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
On Wed, 2009-12-09 at 11:34 -0600, lars_nor...@mcafee.com wrote:
>
> > -Original Message-
> > From: Shawn [mailto:sh...@artemide.us]
> > Sent: Wednesday, December 09, 2009 12:13 PM
> > To: bacula-users@lists.sourceforge.net
> > Subject: Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole
> > is hung up.
> >
> > Hello again, folk,
> >
> > Checked out there and there's really only a few negligible results
> > on this problem.
> >
> > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the
> > bconsole program just won't let me connect.
> >
> > I've checked both .conf files, passwords/addresses/etc, tried
> > changing address on both to name, IP, etc. but it won't let me connect.
> >
> > Output of:
> >
> > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt
> >
> > Connecting to Director 10.6.4.238:9101
> > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip
> > address hidden):9101] All host[ipv4:(ip address hidden):9101]
> > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon
> > host=(ip address hidden) port=9101
> > (Gets hung here for about 300 seconds or so)
> > Director authorization problem.
> > Most likely the passwords do not agree.
> > If you are using TLS, there may have been a certificate validation
> > error during the TLS handshake.
> > Please see http://www.bacula.org/en/rel-
> > manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for
> > help.
> >
> >
> > All I see that's noticeable, while debugging bacula-dir, is that
> > the client tries to connect, and that's about it:
> >
> > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client
> > host=(ip address hidden) port=36131
> >
> > I'm actually not using TLS or anything, has anyone encountered this
> > oddity before?
> >
> > Thanks in advance,
>
>
> Have you done a packet trace on the bacula server? May be (guessing here) as
> a part of the upgrade that host-based firewalling got enabled?
>
> --
> Return on Information:
> Google Enterprise Search pays you back
> Get the facts.
> http://p.sf.net/sfu/google-dev2dev
> ___
> Bacula-users mailing list
> Bacula-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bacula-users
Lars,
Well, truth be told, I was doing this locally on the server. For
sake of experimentation, though, I ensured bconsole.conf was proper on
my desktop, and ran bconsole while doing a trace.
All that happens, is five packets:
47902 > bacula-dir [SYN]
Server acknowledges:
bacula-dir > 47902 [SYN, ACK]
47902 > bacula-dir [ACK]
Client responds with call to director:
47902 > bacula-dir [PSH, ACK]
00 24 81 f8 08 08 00 1d 92 b6 98 70 08 00 45
00 .$.. ...p..E.
0010 00 52 c7 a5 40 00 40 06 55 18 0a 06 04 ef 0a 06 @.@.
U...
0020 04 ee bb 1e 23 8d 86 8e 1c 49 43 cc 0b ef 80
18 #... .IC.
0030 00 2e 4b 5d 00 00 01 01 08 0a 00 76 db eb 0f
6e ..K] ...v...n
0040 7b cc 00 00 00 1a 48 65 6c 6c 6f 20 2a 55 73 65 {.He llo
*Use
0050 72 41 67 65 6e 74 2a 20 63 61 6c 6c 69 6e 67 0a rAgent*
calling.
Server Acknowledges...then, nothing happens at all after this:
bacula-dir > 47902 [ACK]
00 1d 92 b6 98 70 00 24 81 f8 08 08 08 00 45
00 .p.$ ..E.
0010 00 34 70 23 40 00 40 06 ac b8 0a 06 04 ee 0a
06 @.@.
0020 04 ef 23 8d bb 1e 43 cc 0b ef 86 8e 1c 67 80
10 ..#...C. .g..
0030 00 2e 1f ae 00 00 01 01 08 0a 0f 6e 7b cc 00
76 ...n{..v
0040 db
eb ..
So what's puzzling here, is that the server is listening and even
responding, yet it seems like it's hung on the response for some reason.
That, or perhaps there's some communication problem in the package of
bacula-console I'm working with, both in ressu's 3.0.2 and the original
2.4.4, with this director.
I may have to go back to 2.4.4 sadly, if I can't fix this - which
would bring several other bug investigations back to a standstill. I am
currently using ressu's PPA on this .deb package. I will try to switch
to the current unstable Ubuntu development repo's, see if maybe that
package is happier, but, I still can't resist the urge to find out
what's really going on here...
--
Shawn Qureshi
Artemide, Inc.
IT Specialist
<>--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Steve, I do not have any encryption turned on unless it is on by default. I know I did not compile SSL into Bacula. I have not looked at top for long periods but it is a dual quad core system with 4 gigs of ram. I have not seen any of the cores maxed out. Thanks -H -Original Message- From: Steve Polyack [mailto:kor...@comcast.net] Sent: Wednesday, December 09, 2009 10:37 AM To: Hayden Katzenellenbogen Cc: mehma sarja; bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed Hayden Katzenellenbogen wrote: > Everyone, > > Here is a little update. I finally after many months have been able to > run my first restore. > > I started an 800G restore. Which is half of my 1.7TB backup. > > I ran iostat while it was restoring and was getting between 50 and 140 > Meg/s. I would say the average is around 85Meg/s. > > This is performing a restore to the raid-6 array. Not sure if this > information helps or hinders the issue at hand. > > Thanks again for all the good suggestions. > > -H Do you have data encryption enabled? Also, I notice you've mentioned checking iostat during your backups, but have you watched top(1) to monitor CPU usage during the backup? -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Hayden Katzenellenbogen wrote: > Everyone, > > Here is a little update. I finally after many months have been able to > run my first restore. > > I started an 800G restore. Which is half of my 1.7TB backup. > > I ran iostat while it was restoring and was getting between 50 and 140 > Meg/s. I would say the average is around 85Meg/s. > > This is performing a restore to the raid-6 array. Not sure if this > information helps or hinders the issue at hand. > > Thanks again for all the good suggestions. > > -H Do you have data encryption enabled? Also, I notice you've mentioned checking iostat during your backups, but have you watched top(1) to monitor CPU usage during the backup? -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Everyone, Here is a little update. I finally after many months have been able to run my first restore. I started an 800G restore. Which is half of my 1.7TB backup. I ran iostat while it was restoring and was getting between 50 and 140 Meg/s. I would say the average is around 85Meg/s. This is performing a restore to the raid-6 array. Not sure if this information helps or hinders the issue at hand. Thanks again for all the good suggestions. -H From: Hayden Katzenellenbogen [mailto:hay...@nextlevelinternet.com] Sent: Wednesday, December 09, 2009 8:35 AM To: mehma sarja; bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed Mehma, Sorry man wish that was the case. It was doing this on 3.0.1 and then I decided that maybe an upgrade would help. It was doing this when the database was on the local machine. I moved it to its own machine and it was still slow. I have also upgraded the kernel. I will get our db admin to restart the postgres process and see if it makes any difference. I will try anything :) Thanks for the suggestion. -H From: mehma sarja [mailto:mehmasa...@gmail.com] Sent: Tuesday, December 08, 2009 8:52 PM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed Ah HAH! The old "I just upgraded to bacula 3.03 this morning" mistake. The problem here could be what happened to me. The database needs to be bounced. Yes, indexes dropped and re-init-ed. It was a few moons ago and I forget what exactly I did. Horribly slow backups and they NEVER finished and no-one on the list could figure it out. I did the batch thing and pulled my hair out. Until I just happen to bounce PostgreSQL db and voila everything started working. PLEASE BE CAREFUL: a. I do know I re-init-ed the indexes - I had to dump and reload the database and I re-init-ed it in the process. b. This is a delicate process and I played with it for many weeks. Mehma -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
> 2009/12/9 Shawn > > > > Hello again, folk, > > > > Checked out there and there's really only a few negligible results on > > this problem. > > > > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the > > bconsole program just won't let me connect. > > > > I've checked both .conf files, passwords/addresses/etc, tried changing > > address on both to name, IP, etc. but it won't let me connect. > > > > Output of: > > > > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt > > > > Connecting to Director 10.6.4.238:9101 > > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip address > > hidden):9101] All host[ipv4:(ip address hidden):9101] > > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon host=(ip > > address hidden) port=9101 > > (Gets hung here for about 300 seconds or so) > > Director authorization problem. > > Most likely the passwords do not agree. > > If you are using TLS, there may have been a certificate validation error > > during the TLS handshake. > > Please see > > http://www.bacula.org/en/rel-manual/Bacula_Freque_Asked_Questi.html#SECTION00376 > > for help. > > > > > > All I see that's noticeable, while debugging bacula-dir, is that the > > client tries to connect, and that's about it: > > > > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client host=(ip > > address hidden) port=36131 > > > > I'm actually not using TLS or anything, has anyone encountered this > > oddity before? > > > > Thanks in advance, > > > > Did you verify that the director is actually started? > > > -- > John M. Drescher Yes, however, I don't even think clients are able to connect via bacula-fd, as I don't see the backup directory expanding in size or modifying any media/volumes (backup-to-disk) - I know iptables isn't running on this specific server, though, I believe apparmor might be: ps -e | grep bacu 19118 ?00:00:00 bacula-sd 19132 ?00:00:00 bacula-fd 25100 ?00:00:00 bacula-dir I think it's just not listening properly for any of the services...yet, I'm not sure what went wrong here: netstat -la | grep bac tcp0 0 (hostnamehidden):bacula-dir *:* LISTEN tcp0 0 *:bacula-fd *:* LISTEN tcp0 0 *:bacula-sd *:* LISTEN tcp 31 0 (hostnamehidden):bacula-dir (hosthidden):36094 CLOSE_WAIT -- Shawn Qureshi Artemide, Inc. IT Specialist 1980 New Hwy Farmingdale, NY 11735 ph. 631.694.9292 <>-- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
> -Original Message- > From: Shawn [mailto:sh...@artemide.us] > Sent: Wednesday, December 09, 2009 12:13 PM > To: bacula-users@lists.sourceforge.net > Subject: Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole > is hung up. > > Hello again, folk, > > Checked out there and there's really only a few negligible results > on this problem. > > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the > bconsole program just won't let me connect. > > I've checked both .conf files, passwords/addresses/etc, tried > changing address on both to name, IP, etc. but it won't let me connect. > > Output of: > > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt > > Connecting to Director 10.6.4.238:9101 > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip > address hidden):9101] All host[ipv4:(ip address hidden):9101] > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon > host=(ip address hidden) port=9101 > (Gets hung here for about 300 seconds or so) > Director authorization problem. > Most likely the passwords do not agree. > If you are using TLS, there may have been a certificate validation > error during the TLS handshake. > Please see http://www.bacula.org/en/rel- > manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for > help. > > > All I see that's noticeable, while debugging bacula-dir, is that > the client tries to connect, and that's about it: > > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client > host=(ip address hidden) port=36131 > > I'm actually not using TLS or anything, has anyone encountered this > oddity before? > > Thanks in advance, Have you done a packet trace on the bacula server? May be (guessing here) as a part of the upgrade that host-based firewalling got enabled? -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
2009/12/9 Shawn > > Hello again, folk, > > Checked out there and there's really only a few negligible results on > this problem. > > Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the > bconsole program just won't let me connect. > > I've checked both .conf files, passwords/addresses/etc, tried changing > address on both to name, IP, etc. but it won't let me connect. > > Output of: > > bconsole -c /etc/bacula/bconsole.conf -d 100 -dt > > Connecting to Director 10.6.4.238:9101 > 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip address > hidden):9101] All host[ipv4:(ip address hidden):9101] > 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon host=(ip > address hidden) port=9101 > (Gets hung here for about 300 seconds or so) > Director authorization problem. > Most likely the passwords do not agree. > If you are using TLS, there may have been a certificate validation error > during the TLS handshake. > Please see > http://www.bacula.org/en/rel-manual/Bacula_Freque_Asked_Questi.html#SECTION00376 > for help. > > > All I see that's noticeable, while debugging bacula-dir, is that the > client tries to connect, and that's about it: > > 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client host=(ip > address hidden) port=36131 > > I'm actually not using TLS or anything, has anyone encountered this > oddity before? > > Thanks in advance, > Did you verify that the director is actually started? -- John M. Drescher -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Post-upgrade, from 2.4.4 to 3.0.2, bconsole is hung up.
Hello again, folk, Checked out there and there's really only a few negligible results on this problem. Following a full upgrade, from 2.4.4 to 3.0.2 (all packages), the bconsole program just won't let me connect. I've checked both .conf files, passwords/addresses/etc, tried changing address on both to name, IP, etc. but it won't let me connect. Output of: bconsole -c /etc/bacula/bconsole.conf -d 100 -dt Connecting to Director 10.6.4.238:9101 09-Dec-2009 11:46:56 bconsole: bsock.c:221-0 Current host[ipv4:(ip address hidden):9101] All host[ipv4:(ip address hidden):9101] 09-Dec-2009 11:46:56 bconsole: bsock.c:155-0 who=Director daemon host=(ip address hidden) port=9101 (Gets hung here for about 300 seconds or so) Director authorization problem. Most likely the passwords do not agree. If you are using TLS, there may have been a certificate validation error during the TLS handshake. Please see http://www.bacula.org/en/rel-manual/Bacula_Freque_Asked_Questi.html#SECTION00376 for help. All I see that's noticeable, while debugging bacula-dir, is that the client tries to connect, and that's about it: 09-Dec-2009 11:46:56 (dir host hidden): bnet.c:670-0 who=client host=(ip address hidden) port=36131 I'm actually not using TLS or anything, has anyone encountered this oddity before? Thanks in advance, -- Shawn Qureshi Artemide, Inc. IT Specialist <>-- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Mehma, Sorry man wish that was the case. It was doing this on 3.0.1 and then I decided that maybe an upgrade would help. It was doing this when the database was on the local machine. I moved it to its own machine and it was still slow. I have also upgraded the kernel. I will get our db admin to restart the postgres process and see if it makes any difference. I will try anything :) Thanks for the suggestion. -H From: mehma sarja [mailto:mehmasa...@gmail.com] Sent: Tuesday, December 08, 2009 8:52 PM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed Ah HAH! The old "I just upgraded to bacula 3.03 this morning" mistake. The problem here could be what happened to me. The database needs to be bounced. Yes, indexes dropped and re-init-ed. It was a few moons ago and I forget what exactly I did. Horribly slow backups and they NEVER finished and no-one on the list could figure it out. I did the batch thing and pulled my hair out. Until I just happen to bounce PostgreSQL db and voila everything started working. PLEASE BE CAREFUL: a. I do know I re-init-ed the indexes - I had to dump and reload the database and I re-init-ed it in the process. b. This is a delicate process and I played with it for many weeks. Mehma -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Yup got one of those. It has saved my bacon once already. -H -Original Message- From: Dan Langille [mailto:d...@langille.org] Sent: Wednesday, December 09, 2009 4:03 AM To: Hayden Katzenellenbogen Cc: Timo Neuvonen; bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed Hayden Katzenellenbogen wrote: > TiN, > > The write-cache is already enabled. Just checked. I have turned all the settings onto performance just in case. But looking at iostat it does not seem to have made any difference. Write cache should only be enabled if the card has a BBU (battery backup unit). Without a BBU, the data in the cache will be lost should power be lost. -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Slow backup rate
Sean M Clark wrote:
> Carlo Filippetto wrote:
>
>> Hi,
>> I would like to know if is true that I have so slow troughput as this:
>>
> [...]
>
>> FULL
>> -
>> Elapsed time: 1 day 22 hours 13 mins 37 secs
>>
> [...]
>
>> Rate: 371.7 KB/s
>> Software Compression: 15.5 %
>>
>>
> [...]
>
>> All my jobs have the maximum compression
>>
>> /Options {
>> compression = GZIP9 #aggiungo compressione massima
>>
>>
>> /P.S. is this the maximum compression??
>>
> I saw someone else mentioned this already, but yes, GZIP9 is the
> maximum, and that might actually be WHY the rate is slow.
> The higher you set the compression rate, the more time the program
> spends trying to cram more data into each packet before sending it.
> If the compression rate is high enough, it may actually take much
> more time to do the compression than is saved by sending less data.
>
>
I can attest to this, GZIP compression (especially at GZIP9) can make a
large difference in backup transfer speeds, especially when combined
with client data encryption. I wouldn't be surprised if he was to get
5-10MB/sec after disabling GZIP compression, especially if the client
does not have modern hardware.
> Unless space on the backup media or bandwidth usage are the biggest
> concerns, I tend to drop the compression all the way down to
> GZIP1-GZIP4, or turn it off altogether.
>
This is very true. There is often very little gain in compression ratio
with levels greater than GZIP1 or GZIP2, while higher levels will use
considerably more CPU time. Of course this depends entirely on the data
you are compressing, so he should try the various levels himself to see
whether there is any advantage to using a level greater than GZIP1.
> ( I'm hoping to someday see "LZMA1-LZMA9" or "XZ1-XZ9" compression
> options, too... )
>
>
Ditto.
-Steve
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Steve, 1. I am seeing this performance when doing full backups. 2. I am not using software compression as the TL2000 provides hardware compression by default 3. The data is 1.7TB and is around 750 000 files. The actual directories have around 2.1 million files. (using regex to only backup the .wav files) 4. I have done a straight copy between the raid-6 and the spool drive and get around 50Meg/s -H -Original Message- From: Steve Ellis [mailto:el...@brouhaha.com] Sent: Tuesday, December 08, 2009 4:33 PM To: bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] Network transfer Speed I have what sounds like it is a less-powerful system than yours, and I see significantly faster performance from Bacula 3.0.2 (and before that with 2.4 and earlier). My system uses a 3ware 9500 connected merely via 32-bit PCI, and I have a single separate spool drive connected via the motherboards SATA ports. My machine does have motherboard GigE, but as previously pointed out that shouldn't matter. Director, File Daemon, Storage Daemon and MySQL all reside on the same system. Just to reconfirm: * Is this the performance you see with full backups? (differentials and incrementals will always be slower, of course). * Are you avoiding using software compression? * Are the files from the raid array very small or with thousands of files per directory? (which can be important depending on the filesystem used). You may want to verify that the spool drive is performing as expected by just doing some copying from the raid array to the spool drive--although I doubt that is the issue you are hitting. -se On 12/8/2009 2:38 PM, Hayden Katzenellenbogen wrote: > Write cache is enabled. > > I have a separate sata drive connected to the motherboards sata controller > that I am using to spool the data. > > The only thing that is coming off this raid array is the data to be backed > up. Writing to that array is done only once per hour and totals about 1gig > per day. > > I must say I am at a total loss. > > -H > > -Original Message- > From: John Drescher [mailto:dresche...@gmail.com] > Sent: Tuesday, December 08, 2009 2:30 PM > To: Timo Neuvonen; bacula-users > Subject: Re: [Bacula-users] Network transfer Speed > > On Tue, Dec 8, 2009 at 5:13 PM, Timo Neuvonen wrote: > >> "John Drescher" kirjoitti viestissä >> news:387ee2020912081330v6e8fa197s5009971859acb...@mail.gmail.com... >> >>> On Tue, Dec 8, 2009 at 4:16 PM, Hayden Katzenellenbogen >>> wrote: >>> Yes it is a raid-6 configuration running on a 3ware 9690SA-8I. >> I've never had this card, it should be powerful one I think. But I've >> sometimes experienced really poor write performance with 3ware 9550 and 9650 >> cards when unit's write cache was disabled (controller default). If it is >> disabled, try if it has any effect. >> >> > Yes if the cache is disabled you will get horrible random write > performance. If this is also the spool drive that can be the reason. > > John > > -- > Return on Information: > Google Enterprise Search pays you back > Get the facts. > http://p.sf.net/sfu/google-dev2dev > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > > -- > Return on Information: > Google Enterprise Search pays you back > Get the facts. > http://p.sf.net/sfu/google-dev2dev > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Slow backup rate
Carlo Filippetto wrote:
> Hi,
> I would like to know if is true that I have so slow troughput as this:
[...]
> FULL
> -
> Elapsed time: 1 day 22 hours 13 mins 37 secs
[...]
> Rate: 371.7 KB/s
> Software Compression: 15.5 %
>
[...]
> All my jobs have the maximum compression
>
> /Options {
> compression = GZIP9 #aggiungo compressione massima
>
>
> /P.S. is this the maximum compression??
You didn't mention what speed your network connection is, but that's
slow even for 10BaseT. I would expect to see at least about 800 KB/s on
10Mbit, or at least 6-8MB/s on 100MB, or at least 25-35MB/s on Gigabit
(Yeah, I know Gigabit ought to be up around 80MB/s at least, but in
practice I've yet to see anything manage it personally. I know it's
possible, though.)
I saw someone else mentioned this already, but yes, GZIP9 is the
maximum, and that might actually be WHY the rate is slow.
The higher you set the compression rate, the more time the program
spends trying to cram more data into each packet before sending it.
If the compression rate is high enough, it may actually take much
more time to do the compression than is saved by sending less data.
A stupid analogy: if bacula-fd is the shipping department of a company,
"no compression" means the stuff (file data) being shipped is dumped
into a box until it reaches the top, then the box is closed and sent on
its way. Compression level 1 would be like pausing to press down on the
stuff in the box once and then top off the extra space with a little bit
more file data before sending it. Compression level 9 is like dumping
the stuff in, smashing it down, dumping more in, smashing it down,
dumping more in, jumping up and down on top of it, then recruiting some
guys from the next department over to stand on top while you seal the
box. The box ends up holding a lot more, but it takes so much longer to
get the box ready to go that you end up not getting as much shipped out
in the same amount of time.
It can be even worse if the client machine is comparatively low in CPU
power or is heavily loaded (e.g. an old Windows box running Symantec
antivirus doing "active protection" and scanning every file that bacula
tries to examine or send...).
Unless space on the backup media or bandwidth usage are the biggest
concerns, I tend to drop the compression all the way down to
GZIP1-GZIP4, or turn it off altogether.
On the other hand (or "other thread", as the case may be, looking at the
discussion of bandwidth throttling), setting an unnecessarily high
compression level might also be used as a crude way of limiting
bandwidth usage if you don't care so much how long the backup actually
takes.
( I'm hoping to someday see "LZMA1-LZMA9" or "XZ1-XZ9" compression
options, too... )
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Windows 2008 - Can't see restore files
"Carlo Filippetto" kirjoitti viestissä news:8791c1920912090626o4735d964v7eb1031c885a0...@mail.gmail.com... > Hi, > I'm testing to backup two Windows 2008, the first 32 bit, the second 64bit > > In both case the backup jobs go well, but if I tried to 'restore' the > files I can't see them as I post here: > > > - > Automatically selected FileSet: Windows-bacula > +---+---+--++-++ > | JobId | Level | JobFiles | JobBytes | StartTime | VolumeName > | > +---+---+--++-++ > |21 | F |3 | 14,150,553 | 2009-12-09 15:04:33 | > VOL-daily-0004 | > +---+---+--++-++ > You have selected the following JobId: 21 > > Building directory tree for JobId(s) 21 ... > 2 files inserted into the tree. > > You are now entering file selection mode where you add (mark) and > remove (unmark) files to be restored. No files are initially added, unless > you used the "all" keyword on the command line. > Enter "done" to leave this mode. > > cwd is: / > $ ls > C:/ > $ dir > -- 0 root root 0 1970-01-01 01:00:00 C:/ > $ mark > No files marked. > $ mark * > 4 files marked. > $ exit > -- > > As you can read if I ask to list the directory I can't see nothing, and > show the date of 1970-01-01, > If I mark '*' bacula marks 4 files.. and the restore is good! > WHY?? > What can I do? cwd is: / $ ls D:/ $ cd D:/ cwd is: D:/ $ ls Download/ $ mark Download 35 files marked. $ Shortly: add the file/directory name after "mark". Or add the wildcard *, to mark all. Or "cd" to a lower level directory first and... Then finish with "done", not "exit" (maybe it works too, I don't know) -- TiN -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Windows 2008 - Can't see restore files
Hi, I'm testing to backup two Windows 2008, the first 32 bit, the second 64bit In both case the backup jobs go well, but if I tried to 'restore' the files I can't see them as I post here: - Automatically selected FileSet: Windows-bacula +---+---+--++-++ | JobId | Level | JobFiles | JobBytes | StartTime | VolumeName | +---+---+--++-++ |21 | F |3 | 14,150,553 | 2009-12-09 15:04:33 | VOL-daily-0004 | +---+---+--++-++ You have selected the following JobId: 21 Building directory tree for JobId(s) 21 ... 2 files inserted into the tree. You are now entering file selection mode where you add (mark) and remove (unmark) files to be restored. No files are initially added, unless you used the "all" keyword on the command line. Enter "done" to leave this mode. cwd is: / $ ls C:/ $ dir -- 0 root root 0 1970-01-01 01:00:00 C:/ $ mark No files marked. $ mark * 4 files marked. $ exit -- As you can read if I ask to list the directory I can't see nothing, and show the date of 1970-01-01, If I mark '*' bacula marks 4 files.. and the restore is good! WHY?? What can I do? Thank's -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
Hayden Katzenellenbogen wrote: > TiN, > > The write-cache is already enabled. Just checked. I have turned all the > settings onto performance just in case. But looking at iostat it does not > seem to have made any difference. Write cache should only be enabled if the card has a BBU (battery backup unit). Without a BBU, the data in the cache will be lost should power be lost. -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Rate limiting in bacula?
Anyone can recommend a application for Microsoft Windows to limit the bandwidth? Looking for something that is as invisible as possible for the user and only limit the bacula-fd process or port. Freeware or open source of course.. On 12/09/2009 11:14 AM, Thomas Mueller wrote: > >> >> Is it possible to rate limit how fast a backup is sent to the SD? I have >> three collocated boxes, and they all send data at about 100m/s! We're on >> a 20 meg line, so you can imagine what happens to the rest of our >> internet abilities when the backups are running. Any ideas on how to >> slow things down to a manageable pace? >> > use trickle on the filedaemon. > http://monkey.org/~marius/pages/?page=trickle > > > - Thomas > > > -- > Return on Information: > Google Enterprise Search pays you back > Get the facts. > http://p.sf.net/sfu/google-dev2dev > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Network transfer Speed
mehma sarja wrote: > PLEASE BE CAREFUL: > a. I do know I re-init-ed the indexes - I had to dump and reload the > database and I re-init-ed it in the process. > b. This is a delicate process and I played with it for many weeks. This can be achieved without dropping the database. Indexes can be recreated while the database remains intact. If you want to know more, please start a new thread. :) -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Fatal append error on device - Command error with FD
I have tried to add Heartbeat Interval = 120 on both fd and dir. Don't know if it should also be added to SD? -Oprindelig meddelelse- Fra: John Drescher [mailto:dresche...@gmail.com] Sendt: 8. december 2009 15:04 Til: Ralf Gross; bacula-users Emne: Re: [Bacula-users] Fatal append error on device - Command error with FD >> 01-Dec 23:06 kopi3.domain.dk-dir JobId 44: Start Backup JobId 44, >> Job=aspjob.2009-12-01_23.05.00_05 01-Dec 23:06 kopi3.domain.dk-dir JobId >> 44: Max configured use duration exceeded. Marking Volume "Vol0014" as >> Used. >> 02-Dec 01:22 kopi3.domain.dk-dir JobId 44: Fatal error: Network >> error with FD during Backup: ERR=Operation timed out 02-Dec 01:22 >> kopi3.domain.dk-sd JobId 44: Job aspjob.2009-12-01_23.05.00_05 marked to >> be canceled. > > > it's a bit hard to read your mail with all the extra characters in > there... > I believe that is caused by the forum. John -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Slow backup rate
Hi Carlo,
for any modern hardware your rates sound low.
Below is an example I get in my home system (Core2 Duo, 8GB memory, CentOS
5.4 Linux 64-bit), writing to external USB disk, with no compression.
Backing up a local disk, catalog database on the same physical disk too (not
an ideal combination).
FD Files Written: 194,837
SD Files Written: 194,837
FD Bytes Written: 164,511,043,989 (164.5 GB)
SD Bytes Written: 164,537,630,363 (164.5 GB)
Rate: 24175.0 KB/s
Software Compression: None
AFAIK, the GZIP9 you are using is the heaviest compression, both in terms of
expected ratio, and required cpu load. I think the Bacula documentation
mentions that levels over 6 usually result in no significant improvent in
ratio, but consume more cpu power.
Incremental will be slower than full anyway. But since it's this much slower
even in this speed class, makes me think the reason might be something else
than compression. But since it shows very different compression ratio, it
may also be because of compression / different type of the contents of the
files included in the job in average.
At first, try lowering the compression level, or totally disable it, to get
a reference that helps you restrict the possible reasons to the low
throughtput.
A guess without knowing your system: if you have a Windows client with
antivirus sw that handles every disk access, it could have a heavy impact on
this too.
--
TiN
"Carlo Filippetto" kirjoitti viestissä
news:8791c1920912090138l3afdd208t8ec71c4678b1f...@mail.gmail.com...
Hi,
I would like to know if is true that I have so slow troughput as this:
CATALOG
---
FD Bytes Written: 478,808,703 (478.8 MB)
SD Bytes Written: 478,809,069 (478.8 MB)
Rate: 402.0 KB/s
Software Compression: None
INCREMENTAL
--
SD Bytes Written: 40,582,899 (40.58 MB)
Rate: 129.7 KB/s
Software Compression: 86.2 %
SD Bytes Written: 32,037,212 (32.03 MB)
Rate: 179.9 KB/s
Software Compression: 89.7 %
FULL
-
Elapsed time: 1 day 22 hours 13 mins 37 secs
Priority: 10
FD Files Written: 237,200
SD Files Written: 237,200
FD Bytes Written: 61,851,118,685 (61.85 GB)
SD Bytes Written: 61,883,017,775 (61.88 GB)
Rate: 371.7 KB/s
Software Compression: 15.5 %
All my jobs have the maximum compression
Options {
compression = GZIP9 #aggiungo compressione massima
P.S. is this the maximum compression??
My backups are made on Hard Disk. I tried to use an USB, as an iscsi, as a
e-sata device but the rate is so slow..
My best throughput is less then 2Mb/s
It's all right, or there something wrong?
Thak's
Carlo (Italy)
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Rate limiting in bacula?
> > Is it possible to rate limit how fast a backup is sent to the SD? I have > three collocated boxes, and they all send data at about 100m/s! We're on > a 20 meg line, so you can imagine what happens to the rest of our > internet abilities when the backups are running. Any ideas on how to > slow things down to a manageable pace? use trickle on the filedaemon. http://monkey.org/~marius/pages/?page=trickle - Thomas -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Rate limiting in bacula?
Hi, On Wed, 09 Dec 2009, Gabriel - IP Guys wrote: > Is it possible to rate limit how fast a backup is sent to the SD? I have > three collocated boxes, and they all send data at about 100m/s! We're on > a 20 meg line, so you can imagine what happens to the rest of our > internet abilities when the backups are running. Any ideas on how to > slow things down to a manageable pace? I'm not sure if Bacula has such a facility (I haven't noticed one), but a couple of options spring to mind: 1. If you have a firewall in between, can you throttle transfers on tcp port 9103? 2. If your FD runs linux, you could load and use the tcp_lp congestion control module. I haven't tested it particularly but the principal is that it works as a "bandwidth scavenger". It works out the minimum round trip time on a link and wherever it sees the real RTT go higher than that (indicating congestion) it backs off the congestion window, which slows down the transfer to relieve the congestion. Any other connection which needs the link should "win" the bandwidth in that instance. gavi...@ceartgoleor:~$ modinfo /lib/modules/2.6.31-16-generic/kernel/net/ipv4/tcp_lp.ko filename: /lib/modules/2.6.31-16-generic/kernel/net/ipv4/tcp_lp.ko description:TCP Low Priority license:GPL author: Wong Hoi Sing Edison, Hung Hing Lun Mike srcversion: 8BFC408F81AB96C2D21A317 depends: vermagic: 2.6.31-16-generic SMP mod_unload modversions 586 3. There are other tricks you can play such as trickle: http://www.tuxradar.com/content/control-your-bandwidth-trickle I guess you could put that in the init script for either the SD or FD, though it would affect all transfers for that instance which might not be what you want. Gavin -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Slow backup rate
Hi,
I would like to know if is true that I have so slow troughput as this:
*CATALOG
---
** FD Bytes Written: 478,808,703 (478.8 MB)
SD Bytes Written: 478,809,069 (478.8 MB)
Rate: 402.0 KB/s
Software Compression: None
INCREMENTAL
--
SD Bytes Written: 40,582,899 (40.58 MB)
Rate: 129.7 KB/s
Software Compression: 86.2 %
SD Bytes Written: 32,037,212 (32.03
MB)
Rate: 179.9
KB/s
Software Compression: 89.7 %
FULL
-
Elapsed time: 1 day 22 hours 13 mins 37 secs
Priority: 10
FD Files Written: 237,200
SD Files Written: 237,200
FD Bytes Written: 61,851,118,685 (61.85 GB)
SD Bytes Written: 61,883,017,775 (61.88 GB)
Rate: 371.7 KB/s
Software Compression: 15.5 %
*
All my jobs have the maximum compression
*Options {
compression = GZIP9 #aggiungo compressione massima
*P.S. is this the maximum compression??
My backups are made on Hard Disk. I tried to use an USB, as an iscsi, as a
e-sata device but the rate is so slow..
My best throughput is less then 2Mb/s
It's all right, or there something wrong?
Thak's
Carlo (Italy)
--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Rate limiting in bacula?
Dear All, Is it possible to rate limit how fast a backup is sent to the SD? I have three collocated boxes, and they all send data at about 100m/s! We're on a 20 meg line, so you can imagine what happens to the rest of our internet abilities when the backups are running. Any ideas on how to slow things down to a manageable pace? --- Kind Regards, Mr Gabriel -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Possible to merge backups?
* I would like to merge backups with a "move" type of consolidation. Not a "copy". Is it possible? * If the above is not possible, can I choose which backups to be included in a Virtual Backup (Vbackup)? It seems like Vbackup only support to consolidate everything. Like having a chain of F+I+I+I+I and now want to consolidate the two first backups, F+I, so the chain will look like F+I+F+I+I+I. This makes it possible to remove the first F+I. Reason I want this is to reduce disk space usage. -- Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
