On 9/4/2004 4:51 PM, Wes wrote:
> On 9/4/04 2:42 PM, "Jan Wieck" <JanWieck@yahoo.com> wrote:
>
>> Why isn't doing a restore of those reasonable?
>
> Because of the size and time required. Right now, it takes at least 24
> hours, with a full hardware configuration (multi-CPU, 8 disk SCSI RAID,
24 hours to do what? The pg_dumpall, the restore or both? And what media
is used to save the dump? Same disk array, tape, separate disk array on
separate controller or separate archive server? What is the bottleneck?
Planning the backup solution out of the blue, I would have a separate
archive system. This system is connected via gigabit ethernet and has a
storage array large enough to hold at least 2 full dumps. It also has
the tape drive(s) to write off the dumps from the array for storage in
an alternate location. Preferrably, I would have a little more time to
finish the still missing features in Slony-I and take the dumps off a
replica :-)
Since another purpose is to verify if a dump is consistent, the archive
system would have another simple stripe set that can hold the entire
database, plus a ramdisk holding the xlog.
PostgreSQL will most probably not have any kind of MyISAMcheck or
however that utility is called. IIRC RedHat has something that can check
offline for the integrity of a single heap. But I don't know if it is
still under development or if they realized that this is a dead end
because of the offline limitation.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #