> Hi there,
Hello, Keith,
> I was trying to restore a huge data to local harddisk from an offline backup
> harddisk. There are 36,225,746 files and the data size is 1.7TB.
> Steps to restore data:
> 1. Rebuild catalog using bscan
> 2. bconsole > restore > option 3 "Enter list of comma separated JobIds to
> select" > enter {jobid} > mark folder_to_be_restored
> After entered "mark folder_to_be_restored", it took about 28 hours to show $
> prompt and let me proceed to enter "done". Not sure why took so long. Is it
> possible to do some tuning? I could see there is a process "bacula-dir" which
> occupied 4GB memory only. While postgres processes are below 2GB memory. There
> are 7.7GB is free and 3.5GB for cache.
> System configuration:
> OS: 2.6.35.14-106.fc14.x86_64
> Mem: 16G
> Partition1(for OS): RAID1
> Partition2(for data restoration): RAID1
> Bacula: 5.0.3
> Postgresql: 8.4.9-1.fc14.x86_64
> #postgresql.conf
> shared:2GB
> effective_cache_size = 8GB
> maintenance_work_mem=512MB
> work_mem:128MB
> wal_buffers = 8MB
> checkpoint_segments = 64
> checkpoint_timeout = 20min
> checkpoint_completion_target = 0.9
> synchronous_commit = off
I would use Postgresql "synchronous_commit = on". It's the single option (I
tested) that most improves T/s, but other people from this list have deep
database knowledge and may suggest other modifications.
> #sysctl.conf
> kernel.shmmax = 6442450944
> kernel.shmall = 4194304
> Best,
> Keith
Regards,
--
===========================================================================
Heitor Medrado de Faria - LPIC-III | ITIL-F | Bacula Systems Certified
Administrator II
Do you need Bacula training? http://bacula.us/video-classes/
+55 61 8268-4220
Site: http://bacula.us FB: heitor.faria
===========================================================================
------------------------------------------------------------------------------
_______________________________________________
Bacula-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bacula-users
