19.02.2008 10:20, Rathinasamy, Bhaskaran (ext) wrote: > > Hello Arno, > > Many thanks for your reply. > > The below is the output of backup scheduled with default bacula directory in > local machine. > > This is without spooling. > > 18-Feb 16:21 bw1lib1-dir JobId 17: Start Backup JobId 17, > Job=Client1.2008-02-18_16.21.03 > 18-Feb 16:22 bw1lib1-dir JobId 17: Using Device "Drive-1" > 18-Feb 16:22 bw1lib1-sd JobId 17: 3301 Issuing autochanger "loaded? drive 0" > command. > 18-Feb 16:22 bw1lib1-sd JobId 17: 3302 Autochanger "loaded? drive 0", result > is Slot 6. > 18-Feb 16:22 bw1lib1-sd JobId 17: Volume "AAD473" previously written, moving > to end of data. > 18-Feb 16:22 bw1lib1-sd JobId 17: Ready to append to end of Volume "AAD473" > at file=22. > 18-Feb 16:23 bw1lib1-sd JobId 17: Job write elapsed time = 00:00:12, Transfer > rate = 7.673 M bytes/second
Ok, here you've only got a very small amount of data, which makes measuring thoughput more or less useless. The overhead of the tape operations will mask most of the actual data transfer speed. > 18-Feb 16:23 bw1lib1-dir JobId 17: Bacula bw1lib1-dir 2.2.7 (24Dec07): > 18-Feb-2008 16:23:05 > Build OS: i686-pc-linux-gnu suse 9 > JobId: 17 > Job: Client1.2008-02-18_16.21.03 > Backup Level: Full > Client: "bw1lib1-fd" 2.2.7 (24Dec07) > i686-pc-linux-gnu,suse,9 > FileSet: "Full Set" 2008-02-15 18:24:28 > Pool: "Default" (From Job resource) > Storage: "File" (From Job resource) > Scheduled time: 18-Feb-2008 16:21:19 > Start time: 18-Feb-2008 16:22:02 > End time: 18-Feb-2008 16:23:05 > Elapsed time: 1 min 3 secs > Priority: 10 > FD Files Written: 1,812 > SD Files Written: 1,812 > FD Bytes Written: 91,819,734 (91.81 MB) > SD Bytes Written: 92,080,482 (92.08 MB) > Rate: 457.5 KB/s > Software Compression: None > VSS: no > Encryption: no > Volume name(s): AAD473 > Volume Session Id: 1 > Volume Session Time: 1203348070 > Last Volume Bytes: 9,663,704,064 (9.663 GB) > Non-fatal FD errors: 0 > SD Errors: 0 > FD termination status: OK > SD termination status: OK > Termination: Backup OK > > I enabled spooling and the speed is even less, now I tried writing the whole > root directory(/), My spooling directory is /var/backup. Good... regarding spooling, it's normal to decrease the overall throughput of a single job, as each byte of data has to be handled twice by Bacula. > > 19-Feb 09:32 bw1lib1-sd JobId 26: Job write elapsed time = 00:07:08, Transfer > rate = 20.84 M bytes/second > 19-Feb 09:32 bw1lib1-sd JobId 26: Committing spooled data to Volume "AAD473". > Despooling 8,933,523,455 bytes ... > 19-Feb 09:52 bw1lib1-sd JobId 26: Despooling elapsed time = 00:19:23, > Transfer rate = 7.681 M bytes/second But in fact, despooling time is quite long. An LTO drive should be much faster. There are a number of possible reasons, starting with SCSI hardware issues and ending with the file system you read from. Most important is to use btape first and test the drive configuration you have. The test command will give you throughput numbers from memory to tape - these should be considerably higher. If they aren't, it's quite possible you've got SCSI problems or a misconfigured tape drive, kernel, or whatever. If the speed btape measures is decent, it's time to look for bottlenecks on the system. For example, if you're reading from one file system, and your catalog is written to the same disk, this will cost lots of performance. > 19-Feb 09:52 bw1lib1-sd JobId 26: Sending spooled attrs to the Director. > Despooling 33,822,898 bytes ... > 19-Feb 09:52 bw1lib1-dir JobId 26: Bacula bw1lib1-dir 2.2.7 (24Dec07): > 19-Feb-2008 09:52:49 > Build OS: i686-pc-linux-gnu suse 9 > JobId: 26 > Job: Client1.2008-02-19_09.25.03 > Backup Level: Full > Client: "bw1lib1-fd" 2.2.7 (24Dec07) > i686-pc-linux-gnu,suse,9 > FileSet: "Full Set" 2008-02-18 14:49:46 > Pool: "Default" (From Job resource) > Storage: "File" (From Job resource) > Scheduled time: 19-Feb-2008 09:25:22 > Start time: 19-Feb-2008 09:25:26 > End time: 19-Feb-2008 09:52:49 > Elapsed time: 27 mins 23 secs > Priority: 10 > FD Files Written: 124,366 > SD Files Written: 124,366 > FD Bytes Written: 8,904,940,933 (8.904 GB) > SD Bytes Written: 8,921,603,753 (8.921 GB) > Rate: 419.9 KB/s > Software Compression: None > VSS: no > Encryption: no > Volume name(s): AAD473 > Volume Session Id: 1 > Volume Session Time: 1203409443 > Last Volume Bytes: 9,119,416,320 (9.119 GB) > Non-fatal FD errors: 0 > SD Errors: 0 > FD termination status: OK > SD termination status: OK > Termination: Backup OK > > All the backup iam trying is Full and in local machine.And when I was trying > above without spooling the time ( 7 Minutes ) was less. I mean the whole job > got over 7 minutes earlier. > > Any fix or configuration for this.? Or is it the speed bacula can do? Definitely not. Finding the most likely issues, though, requires more knowledge of your setup. Start with btape and its test command, and then let's see if we can pinpoint the problem step by step... Arno > Bhaskaran > -----Original Message----- > From: Arno Lehmann [mailto:[EMAIL PROTECTED] > Sent: Monday, February 18, 2008 8:39 PM > To: Rathinasamy, Bhaskaran (ext) > Subject: Re: [Bacula-users] Hello Help for a Newbie, LTO -3 Bacula 2.2.7 > > Hi, > > 18.02.2008 16:53, Rathinasamy, Bhaskaran (ext) wrote: >> Hello, >> >> Iam trying to setup bacula as below : >> >> OS = SLES 9 with SP3, 4GB memory, 8 GB Swap. >> Version bacula = 2.2.7 >> Tape Library = Fujitsu Siemens TX48.Autochanger with 48 slots. >> Media = LTO -3. >> >> When iam trying to write to the tape drive. The speed is very slow like >> Transfer rate = 7.673 M bytes/second ( This is the nearly native speed >> of the tape drive ) > > How do you measure? I.e. is this the overall throughput from the job > report, btape output, or tape write speed with spooling? > > If it's from an actual job, it is important to know if this is an > incremental or full backup, locally or over the network, and possibly > which sort of file set (many small files vs. few large files). Even > the underlying file system can affect backup speed... > > Arno > > >> I enabled the Hardware compression as given by document, but eventhough >> the speed is 7.673 M bytes/second exactly. >> >> When i use a dump with compression and 1024 Block size, it writes at >> very good speed as below: >> dump -0 -b 1024 -f /dev/nst0 /var >> DUMP: Date of this level 0 dump: Mon Feb 18 16:45:24 2008 >> DUMP: Dumping /dev/mapper/system-SYSTEM_VAR (/var) to /dev/nst0 >> DUMP: Label: none >> DUMP: Writing 1024 Kilobyte records >> DUMP: mapping (Pass I) [regular files] >> DUMP: mapping (Pass II) [directories] >> DUMP: estimated 157674 blocks. >> DUMP: Volume 1 started with block 1 at: Mon Feb 18 16:45:32 2008 >> DUMP: dumping (Pass III) [directories] >> DUMP: 0.65% done, finished in 0:00 >> DUMP: dumping (Pass IV) [regular files] >> DUMP: Closing /dev/nst0 >> DUMP: Volume 1 completed at: Mon Feb 18 16:45:40 2008 >> DUMP: Volume 1 154624 blocks (151.00MB) >> DUMP: Volume 1 took 0:00:08 >> DUMP: Volume 1 transfer rate: 19328 kB/s >> DUMP: 154624 blocks (151.00MB) on 1 volume(s) >> DUMP: finished in 6 seconds, throughput 25770 kBytes/sec >> DUMP: Date of this level 0 dump: Mon Feb 18 16:45:24 2008 >> DUMP: Date this dump completed: Mon Feb 18 16:45:40 2008 >> DUMP: Average transfer rate: 19328 kB/s >> DUMP: DUMP IS DONE >> >> >> But the writing speed with and without Hardware compressiong is exactly >> 7.673 M bytes/second .can someone please help me. Am i wrong in any >> configuration. >> >> >> Thanks and Regards / Mit freundlichen Grüßen, >> >> >> >> Bhaskaran Rathinasamy > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL > PROTECTED] > Sent: Tuesday, February 19, 2008 10:03 AM > To: bacula-users@lists.sourceforge.net > Subject: Bacula-users Digest, Vol 22, Issue 29 > > Send Bacula-users mailing list submissions to > bacula-users@lists.sourceforge.net > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.sourceforge.net/lists/listinfo/bacula-users > or, via email, send a message with subject or body 'help' to > [EMAIL PROTECTED] > > You can reach the person managing the list at > [EMAIL PROTECTED] > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Bacula-users digest..." > > > Today's Topics: > > 1. Re: How do Priorities work across multiple Storages? > (Dan Langille) > 2. 1 Job = many File Sets (Frank Kujawski) > 3. Required Bacula GUI for Configuration ([EMAIL PROTECTED]) > 4. Re: Required Bacula GUI for Configuration (Dan Langille) > 5. Re: 1 Job = many File Sets (Dan Langille) > 6. Re: Required Bacula GUI for Configuration (Bruno Friedmann) > 7. Compiling Bacula in a new system (Reynier Perez Mira) > 8. Hello Help for a Newbie, LTO -3 Bacula 2.2.7 > (Rathinasamy, Bhaskaran (ext)) > 9. Re: Compiling Bacula in a new system (Reynier Perez Mira) > 10. Support for Dell Tape Backup (Reynier Perez Mira) > 11. Dumb tape drive question (Tom Allison) > 12. Re: Dumb tape drive question (Dan Langille) > 13. Re: Dumb tape drive question (John Drescher) > 14. Re: FW: Bacula console configuration failure (Arno Lehmann) > 15. Re: Dumb tape drive question (Tom Allison) > 16. Why Can't I Stat Mapped Vista network drives. > ([EMAIL PROTECTED]) > 17. Re: Why Can't I Stat Mapped Vista network drives. (Frank Sweetser) > 18. Re: Why Can't I Stat Mapped Vista network drives. (John Drescher) > 19. static linked bacula-fd with openssl (Vladimir Vasilev) > 20. Re: static linked bacula-fd with openssl (Bruno Friedmann) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Mon, 18 Feb 2008 07:56:27 -0500 > From: Dan Langille <[EMAIL PROTECTED]> > Subject: Re: [Bacula-users] How do Priorities work across multiple > Storages? > To: James Harper <[EMAIL PROTECTED]> > Cc: bacula users <bacula-users@lists.sourceforge.net> > Message-ID: <[EMAIL PROTECTED]> > Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed > > > On Feb 17, 2008, at 11:03 PM, James Harper wrote: > >>> Unless you are allowing more than one concurrent job. In which case, >>> priority will >>> be your best bet. But I don't recall you mentioning concurrent jobs. >>> >> Hmmm... forgot to mention that :) >> >> Yes, I'm running the jobs concurrently, so I assume that given an >> equal >> priority, whichever one finishes writing to the spool will get >> despooled >> first and finish first, even if it didn't start first? > > I have not worked with concurrent jobs. They are... too complex/ > risky for my > liking. This does not mean I think there is a problem with the > code. I just > prefer not to use them. > > As to your question, I don't know.... it's too early in the day. :0 > > Back to your original question. The subject states "multiple > storages". Is the scenario: two jobs, two storage daemons? > > Or two jobs, one SD? > -- Arno Lehmann IT-Service Lehmann www.its-lehmann.de ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
