Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Peter Sjoberg
On Mon, 2005-04-04 at 07:22 +0200, Michael 'buk' Scherer wrote: > Hey. > > Matt White wrote: > > Below is a backup report - first of all, am I reading this right? It > > looks to me like the actual backup is finished by around 22:42 (data > > actually transferred to asclbackup by 21:15, and then

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Kern Sibbald
On Monday 04 April 2005 18:35, Matt White wrote: > Kern Sibbald wrote: > > Well, there are two obvious ways of speeding up inserting of attributes. > > > > 1. Cache the attributes, and combine them with the MD5/SHA1 signature > > that follows each attribute, and then do a single insert rather than

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Kern Sibbald
On Monday 04 April 2005 18:45, Matt White wrote: > Matt White wrote: > > For PostgreSQL, there is one other thing I can think of that might be > > done to speed things up...single sql statements are executed with an > > implicit BEGIN TRANSACTION/END TRANSACTION. If it's doable without > > affecti

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Kern Sibbald
Hello, One tip I just got from Meno Abels today is to increase the Path index size. For some reason, he didn't have any Path index, eventhough the default should be to have the Path indexed on 50 characters. He has increased it to 512 characters, but on my system, the maximum with MySQL 3.22.53

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Matt White
Matt White wrote: For PostgreSQL, there is one other thing I can think of that might be done to speed things up...single sql statements are executed with an implicit BEGIN TRANSACTION/END TRANSACTION. If it's doable without affecting the other databases, would you be receptive to a patch to wrap t

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Matt White
Kern Sibbald wrote: Well, there are two obvious ways of speeding up inserting of attributes. 1. Cache the attributes, and combine them with the MD5/SHA1 signature that follows each attribute, and then do a single insert rather than an insert of the attribute followed by an update for the MD5/

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Arno Lehmann
Hello. Michael 'buk' Scherer wrote: Hey. Matt White wrote: ... I have the exact same problem, Bacula spools 25GB in ~1.5 million files with a good speed and writes them to tape "in no time". Despooling the attributes takes ages. See the "Speed up backup of small files" thread ~2 weeks ago. The only

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Michael 'buk' Scherer
Alan Brown wrote: How much memory is being allocated for the database server? Uhm, nothing set. As much as it needs I would say? Box has 1GB. What database are you using? MySQL-4.0.18 M. -- I am root. If you see me laughing, you better have a backup. -

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Alan Brown
On Mon, 4 Apr 2005, Michael 'buk' Scherer wrote: I have the exact same problem, Bacula spools 25GB in ~1.5 million files with a good speed and writes them to tape "in no time". Despooling the attributes takes ages. See the "Speed up backup of small files" thread ~2 weeks ago. The only advice I got

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Kern Sibbald
On Monday 04 April 2005 09:55, Michael 'buk' Scherer wrote: > Good day. > > Willing, yes. > Being able to, maybe. > > Details please. :) Well, there are two obvious ways of speeding up inserting of attributes. 1. Cache the attributes, and combine them with the MD5/SHA1 signature that follows e

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Michael 'buk' Scherer
Good day. Willing, yes. Being able to, maybe. Details please. :) M. Kern Sibbald wrote: Hello, Are you willing to do some C programming to get a performance improvement? Best regards, Kern On Monday 04 April 2005 05:42, Matt White wrote: Hi - I've been evaluating Bacula for a while now, and I'm get

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-04 Thread Kern Sibbald
Hello, Are you willing to do some C programming to get a performance improvement? Best regards, Kern On Monday 04 April 2005 05:42, Matt White wrote: > Hi - I've been evaluating Bacula for a while now, and I'm getting set to > transition from Amanda to Bacula. I'm backing up 17 hosts (Linux and

Re: [Bacula-users] Speeding up attribute inserts/updates

2005-04-03 Thread Michael 'buk' Scherer
Hey. Matt White wrote: Below is a backup report - first of all, am I reading this right? It looks to me like the actual backup is finished by around 22:42 (data actually transferred to asclbackup by 21:15, and then another 1.5 hours or so to dump it to tape), and then the next 8 hours are spent de

[Bacula-users] Speeding up attribute inserts/updates

2005-04-03 Thread Matt White
Hi - I've been evaluating Bacula for a while now, and I'm getting set to transition from Amanda to Bacula. I'm backing up 17 hosts (Linux and Tru64 Unix), currently to a 26 tape TL-891 (DLT-IV) library. I've got my new tape drive in (30 tape MSL6030 with an Ultrium-2), and for the last week or so