From: | Michael Nolan <htfoot(at)gmail(dot)com> |
---|---|
To: | Yunong J Xiao <yjxiao(at)gmail(dot)com> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Backups using Solaris ZFS Snapshots |
Date: | 2012-04-24 22:29:51 |
Message-ID: | CAOzAqu+xHb_jyXpSN4nMSnP8aS3zWnG+N4QpZMkvx2BbwW2fRQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Tue, Apr 24, 2012 at 4:08 PM, Yunong J Xiao <yjxiao(at)gmail(dot)com> wrote:
> I am currently backing up my postgres instances using ZFS snapshots
> instead of the sanctioned pg_dump utility mainly because I am running on
> Solaris and it offers a copy-on-write file system. Anecdotally this has
> been working fine for me. Are there any issues I should be aware of since
> I'm not using pg_dumps at all but merely copying snapshots of the postgres
> data directory?
>
As a matter of principle, you should test your backup strategy
periodically, no matter what it is!
This test should tell you:
1. Whether the backup method even works. (Sadly, a backup method that
worked in the past may no longer work.)
2. What steps are needed to recover from a backup.
3. How much data loss (if any) you are likely to experience.
4. How long it will take to bring up the recovered database.
All of these are things that you need to know in advance, and your
management will be vitally interested in #3 and #4 when the occasion arises
to have to use a backup.
You also need to know how to do partial recoveries (such as an
inadvertently deleted or corrupted but important table.) Not all backup
strategies lend themselves readily to partial recoveries.
--
Mike Nolan
From | Date | Subject | |
---|---|---|---|
Next Message | Toby Corkindale | 2012-04-25 06:18:37 | Bug? Query plans / EXPLAIN using gigabytes of memory |
Previous Message | Yunong J Xiao | 2012-04-24 21:08:57 | Backups using Solaris ZFS Snapshots |