xfs
[Top] [All Lists]

Re: Is xfsdump operation atomic?

To: Michael Locher <locher@xxxxxxxxxxxx>
Subject: Re: Is xfsdump operation atomic?
From: Christian Rice <xian@xxxxxxxxxxx>
Date: Wed, 24 Nov 2004 10:43:36 -0800
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <F4C07791-3DF5-11D9-9800-000A95BCCB96@iam.unibe.ch>
Organization: Tippett Studio
References: <38A5E378-3DF4-11D9-9800-000A95BCCB96@iam.unibe.ch> <F4C07791-3DF5-11D9-9800-000A95BCCB96@iam.unibe.ch>
Reply-to: xian@xxxxxxxxxxx
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040113
This is why snapshots were invented.

Michael Locher wrote:
Ivan, thanks for commenting on this.

Am 23.11.2004 um 13:12 schrieb  Ivan Rayner:

You seem to be of the opinion that you can't create a backup of a live
filesystem.  My guess is that you assume a backup is a strict snapshot in
time of the filesystem -- it isn't.


But it could be... it would be a nice feature to have, tough i am sure it is not trivial to implement.

If a file is removed or created while a xfsdump is running, then it may or
may not be included in the dump. This is OK, because in tomorrow night's
incremental backup the file will then be included or removed as
appropriate. If you have a 24hr schedule then you should assume that the
cycle starts when xfsdump starts. Anything that happens while xfsdump is
running will be guaranteed to be on the following dump, but if you're
lucky then it will make it in the current dump.


What happens if a large file (eg a multi GB database file) is modified during the dump.
Will it be consistent? i.e are write operations allowed on a file when its data is dumped?


Michael





<Prev in Thread] Current Thread [Next in Thread>