[Top] [All Lists]


To: linux-xfs@xxxxxxxxxxx
Subject: Re: RHEL ES 4
From: pg_xfs@xxxxxxxxxxxxxxxxxx (Peter Grandi)
Date: Fri, 18 Nov 2005 17:13:46 +0000
In-reply-to: <437E053E.8090704@xxxxxxx>
References: <32927.> <437D6935.2090905@xxxxxxx> <1132326431.12165.9.camel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> <437DFBD8.3070106@xxxxxxx> <437E0297.40807@xxxxxxx> <437E053E.8090704@xxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
>>> On Fri, 18 Nov 2005 17:45:50 +0100, "evilninja@xxxxxxx"
>>> <evilninja@xxxxxxx> said:

evilninja> Eric Sandeen schrieb:

Eric> Oh, repair on a 300T filesystem -wil-l be painful
Eric> anywhere, I think, unfortunately.

evilninja> hm, painful yes, but hopefully not impossible?
evilninja> otherwise if sth. goes wrong on a 300TB fs

I suspect that Eric above is trying to be an master of the

evilninja> the only way to fix it would be restoring from
evilninja> backup, no matter how tiny the corruption might be
evilninja> (and xfs_repair could fix it easliy if only the
evilninja> volume was not so big)....

As an amusing mental exercise, consider this: how to make such
a backup of a 300TB filesystem?

Overall I would guess that designing a 2^48B filesystem is a bit
of a research problem, and requires serious study and thinking,
if one cannot avoid it.

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