SearchFAQMemberlist Log in
Reply to topic Page 1 of 1
bacula reads disk pages atomically? (bdb question)
Author Message
Post bacula reads disk pages atomically? (bdb question) 
The Berkeley Database apparently requires that hot backups read disk
pages atomically. They note that most system cp commands do so; I've
searched the manual, FAQ, and newsgroups, but can't find how bacula
operates when reading a file.

Does anyone know if I can count on the reads taking whole pages?

More generally, any tips on bdb backups with bacula would be great.

I'm on Linux with a 2.6 kernel.

Thanks.
--
Ross Boylan wk: (415) 514-8146
185 Berry St #5700 ross < at > bi...
Dept of Epidemiology and Biostatistics fax: (415) 514-8150
University of California, San Francisco
San Francisco, CA 94107-1739 hm: (415) 550-1062

Post bacula reads disk pages atomically? (bdb question) 
On Tue, 12 Jun 2007 14:14:56 -0700, Ross Boylan said:

The Berkeley Database apparently requires that hot backups read disk
pages atomically. They note that most system cp commands do so; I've
searched the manual, FAQ, and newsgroups, but can't find how bacula
operates when reading a file.

Does anyone know if I can count on the reads taking whole pages?

Generally, no.

Specifically, possibly, at least if the page size is a factor of 65536, you
don't set sparse=yes, the OS guarantees that a call to read(2) with a page
aligned size will do it and also the network allows it. In these cases, it
looks like the current read size equals the network buffer size (see Maximum
Network Buffer Size).

You could try using strace to see what bacula-fd is doing in practice.

__Martin



Re: [Bacula-users] exabyte autochanger From: Maria McKinley

- 2007-06-13 21:47

Broderick Wood wrote:
I would also do a quick check to make sure you have LUN support enabled.



What is LUN support, and how do I check to make sure it is enabled?

~maria


On Wed, 13 Jun 2007 14:24:09 -0700
Michael Nelson <mnelson < at > in...> wrote:

Are you sure that /dev/sg0 is in fact the autochanger device and not the
tape drive?

For instance on my server (from dmesg output):


(scsi3:A:4): 160.000MB/s transfers (80.000MHz DT, offset 127, 16bit)
Vendor: CERTANCE Model: ULTRIUM 3 Rev: 1856
Type: Sequential-Access ANSI SCSI revision: 04
Attached scsi generic sg2 at scsi3, channel 0, id 4, lun 0, type 1

st: Version 20040403, fixed bufsize 32768, s/g segs 256
Attached scsi tape st0 at scsi3, channel 0, id 4, lun 0
st0: try direct i/o: yes (alignment 512 B), max page reachable by HBA
134217727
Vendor: QUANTUM Model: UHDL Rev: 0031
Type: Medium Changer ANSI SCSI revision: 02
Attached scsi generic sg3 at scsi3, channel 0, id 4, lun 1, type 8

So the generic device for my tape drive is /dev/sg2 (but the tape drive
is normally addressed as /dev/nst0) and the changer for it is /dev/sg3.

It would be interesting to see the output of "cat /proc/scsi/scsi" on
your machine.

[root < at > nts ~]# cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: LSILOGIC Model: 1030 IM Rev: 1000
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 08 Lun: 00
Vendor: IBM Model: 25P3495a S320 1 Rev: 1
Type: Processor ANSI SCSI revision: 02
Host: scsi3 Channel: 00 Id: 04 Lun: 00
Vendor: CERTANCE Model: ULTRIUM 3 Rev: 1856
Type: Sequential-Access ANSI SCSI revision: 04
Host: scsi3 Channel: 00 Id: 04 Lun: 01
Vendor: QUANTUM Model: UHDL Rev: 0031
Type: Medium Changer ANSI SCSI revision: 02




--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bacula-users mailing list
Bacula-users < at > li...
https://lists.sourceforge.net/lists/listinfo/bacula-users




Post bacula reads disk pages atomically? (bdb question) 
On Tue, 2007-06-12 at 14:14 -0700, Ross Boylan wrote:
The Berkeley Database apparently requires that hot backups read disk
pages atomically. They note that most system cp commands do so; I've
searched the manual, FAQ, and newsgroups, but can't find how bacula
operates when reading a file.

Does anyone know if I can count on the reads taking whole pages?

More generally, any tips on bdb backups with bacula would be great.

I'm on Linux with a 2.6 kernel.

Thanks.
The bacula manual, under the discussion of FileSet | Sparse says "Bacula
reads files in 32k buffers." I suppose this is not quite definitive
about how the data are actually pulled off the disk.

I don't know what the page size is for Linux (I suppose it might vary by
filesystem), and the page size that individual bdb apps use is
determined by the app (the recommendation is that the app page size
equal the disk page size).
Ross

Display posts from previous:
Reply to topic Page 1 of 1
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
  


Magic SEO URL for phpBB