Gentoo Archives: gentoo-user

From: Pandu Poluan <pandu@××××××.info>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] Heads-up: Several kernel versions have severe EXT4 data corruption bug
Date: Fri, 26 Oct 2012 03:32:07
Message-Id: CAA2qdGW_3-onWsdWstCtMm6ZgQBW7F=nYwQ1orsogVcDKADH-w@mail.gmail.com
In Reply to: Re: [gentoo-user] Heads-up: Several kernel versions have severe EXT4 data corruption bug by Mark Knecht
1 On Oct 26, 2012 3:36 AM, "Mark Knecht" <markknecht@×××××.com> wrote:
2 >
3 > On Thu, Oct 25, 2012 at 12:40 PM, Pandu Poluan <pandu@××××××.info> wrote:
4 > <SNIP>
5 > >
6 > > Nice find, thanks. But the update to that article which gave a link to
7 Theo
8 > > T'so email is quite unnerving. He said that his initial hypothesis had
9 not
10 > > been proven and he himself is back in the dark.
11 > >
12 > > Not much confidence in ext4 for the time being for me. Back to
13 reiserfs? ;-)
14 > >
15 > > Rgds,
16 > > --
17 >
18 > I'm not an ext4 user (it's too new!) and therefore unaffected by any
19 > of this personally. However I do read lkml and have been following the
20 > thread since it started. This bug, if indeed in the end it's even
21 > determined to be a bug, is apparently pretty hard to hit. The downside
22 > is that if you did hit it the effects can be pretty devastating.
23 >
24 > I'd suggest folks consider going through the thread and not
25 > overreacting to the masking of new kernels or stuff posted on Phoronix
26 > or slashdot. It's good that the info is out there so people can make
27 > informed decisions, but one of the downsides of having all the info is
28 > the potential for people to overreact. Even with a ;-) dumping ext4
29 > for Reiserfs might be in that camp. :-)
30 >
31 > Cheers,
32 > Mark
33 >
34
35 LoL... I was just joking about dumping ext4. I still am going to use ext4,
36 but I will ratchet up the backup schedule, and anyone thinking of
37 restarting a Linux server in my office before first performing a full
38 backup, will be severely reprimanded.
39
40 (I *do* use ReiserFS, though. Especially for portage's tmp)
41
42 Rgds,
43 --