xfs-masters
[Top] [All Lists]

[Bug 860] XFS filesystem crash

To: xfs-masters@xxxxxxxxxxx
Subject: [Bug 860] XFS filesystem crash
From: bugzilla-daemon@xxxxxxxxxxx
Date: Fri, 15 Jan 2010 18:38:26 -0600
Auto-submitted: auto-generated
In-reply-to: <bug-860-113@xxxxxxxxxxxxxxxx/bugzilla/>
References: <bug-860-113@xxxxxxxxxxxxxxxx/bugzilla/>
http://oss.sgi.com/bugzilla/show_bug.cgi?id=860





--- Comment #10 from Carlos E. R. M. <carlos.e.r@xxxxxxxxxxxx>  2010-01-15 
18:36:55 CST ---
(In reply to comment #9)
> > Er... excuse the question, what is "volatile write caches"?
> 
> write caches that lose your data when they lose power.
> 
> Consumer hard drives usually have 8-16-32MB write caches; if the drive loses
> power, the cached data evaporates and never hits the disk - but the disk
> already told xfs that the data was safe, and xfs has moved on.  This violation
> of trust between fs and disk can cause filesytem corruption :)

Ah, I understand. I had heard of that, but not of the name.

> Options are to get a ups, or turn off write caches.  Or to find out why you're
> losing those IOs, and keep using barriers.

Mmm. I have an ups, but not yet on this machine. It is in the to-do list. The
data should be safe after a clean umount, but you are right, it is an issue.

The real solution would be for the chaps at Suse/Novell/openSUSE to find out
what is wrong with DM. I have passed this info to them. I'll keep hoping.


My current solution is to use reiserfs instead, for encrypted data at least
>:-)

-- 
Configure bugmail: http://oss.sgi.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

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