bug-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ext3 or other journalling file system for Hurd?


From: olafBuddenhagen
Subject: Re: ext3 or other journalling file system for Hurd?
Date: Tue, 25 Jan 2011 06:34:33 +0100
User-agent: Mutt/1.5.20 (2009-06-14)

Hi,

On Fri, Jan 21, 2011 at 09:39:06PM +0100, Svante Signell wrote:

> (Not serious development though, I have merely an EE background, not
> CS). 

You are mistaken if you seriously believe a formal CS background has
anything to do with this :-)

> I have a problem that remains wrt GNU/Hurd: I'm still getting frequent
> crashes (sometimes due to improper shutdown in qemu), and since only
> the ext2 file system is supported, the file system has to be checked
> at every new boot. Sometimes, even a proper shutdown in qemu results
> in a corrupted file system at next boot.  Recently, I had a serious
> file system corruption and am still suffering from that. Maybe a
> complete reinstall is the only solution at this point. 

That's unusual. I have been running my current installation since 2004
or so, and in spite of frequent minor issues, never had really critical
filesystem damage...

If you have critical FS damage, that's most likely *not* caused by the
unclean shutdown, but something more serious, which resulted in writing
garbage to the disk. Journalling wouldn't help with this at all.

Note that some people have been recently reporting similar FS corruption
issues when running Hurd under KVM -- so this might be the real cause of
your trouble...

> Is there a possibility/how much effort is needed to implement a
> journalling file system like etx3, or other.

Of course it's possible. I'm not sure how much work it is exactly, but
it's certainly not trivial -- and I'm not sure it's really worth the
effort...

> (I think GNU-specific file system would be asking for too much)

I'd love to see one... But there are other priorities obviously :-)

> Does a priority wish list exist for the development,

Nope. Every developer has different priorities...

Having said that, some kind of generic roadmap still would be useful --
but so far nobody got around to compile one :-)

-antrik-



reply via email to

[Prev in Thread] Current Thread [Next in Thread]