another 2.4.12 + aacraid + SuSE failure.

Hendrik Visage hvisage at is.co.za
Fri Nov 2 00:03:00 CST 2001


Using SGL (sorcerer.wox.org) Linux 2.4.13 + patches
It appears to be only for the ext2 FS's and not for the reiserfs.

Interesting, if you've compiled in kernel debugging & the SysReq key stuff, 
and then issue an emergency sync (Alt-SysReq-S) it works to break the ext2
out of the lock. 

An strace of a hanging e2fsck, hangs at the opening of the device...

I'll perhaps have to try the -ac patches then ...

Hendrik

On Wed, Oct 31, 2001 at 12:28:02PM -0600, Steve_Boley at dell.com wrote:
> Been pounding away on recompiling 2.4.10 on PE2500 with PERC3/DI and was
> getting the same hang at CHECKING ROOT FILESYSTEM.  Was trying all the
> different aacraid patches and finally thought "Hey what if it's the
> kernel?????".  So I then decided to go to the kernel/people folder and get
> the latest patch that Alan Cox has for the 2.4.10 kernel which was
> 2.4.10-ac12 and see what happened.  Recompiled with the aacraid patch for
> the kernel and no hang.  Some bug in the base kernel at kernel.org is
> causing the issue which Alan patched in his patches somewhere.  I would
> assume that if you patch 12 and 13 you will get the same results.  I was
> doing this recompile in Mandrake 8.1 and I don't know what their base source
> kernel is but Redhat uses the Alan Cox kernel as it's base kernel and hence
> no problems with redhat recompiles.
> 
> -----Original Message-----
> From: Chris Pascoe [mailto:c.pascoe at itee.uq.edu.au]
> Sent: Thursday, October 25, 2001 8:36 PM
> To: Steve_Boley at exchange.dell.com
> Cc: linux-PowerEdge at exchange.dell.com;
> linux-aacraid-devel at exchange.dell.com
> Subject: Re: another 2.4.12 + aacraid + SuSE failure.
> 
> 
> Hi Steve,
> 
> > This is directed mainly toward Matt and the other guys, almost all
> hardware
> > and software vendors allow people to sign on for beta testing (pretty much
> > what the whole linux experience is) of drivers and firmware.  What about
> > Dell starting a program for our Linux customers to beta drive our drivers
> > and firmware before official release?  Of course this would be with
> > explicitly implied that this would not be the responsibility of Dell for
> > loss of data or any other complications arising from this.  There is a
> newer
> > firmware for the adaptec controllers that will greatly enhance performance
> > over current 2.5 revision that is released.
> 
> I know we have a 4400 identically configured to our production machines,
> (apart from the "rev 01"/"rev 02" issue) for the sole purpose of testing,
> and would be interested in seeing this, for testing as well.  So long as we
> could downgrade again, of course :)
> 
> > You aren't able to go in single mode and manually run a fsck on the
> > partitions in the 2.4.12 kernel?  Looking at kernel lists these kinds of
> > problems were surfacing starting with the 2.4.11 kernel so it's more than
> > likely more a kernel specific type of problem.
> 
> Nope, I couldn't run fsck from single user. I didn't have any other kernels
> handy to boot with other than the 2.4.2 one, so I couldn't try going back to
> 2.4.11.  Perhaps some others can comment on this.  Unfortunately for me,
> it's not an easy thing to replicate - when the filesystem was unmounted
> properly, everything goes through fine, and my few attempts to corrupt it on
> subsequently have been in vain.
> 
> > Have you been monitoring the kernel lists?  There is a wealth of
> information
> > on them but they are quite long and tedious to go through.
> 
> Yes, I try to monitor them; but usually my days run short and I don't get to
> read everything I'd like.  I figured out a way around the 2.4.10 sd.c
> problems through that, and do try to read up.  I didn't see anything
> directly mentioning the fsck problems in my quick searches though.  I
> suspect the lack of time thing hits all of us who can't dedicate time to
> looking through them.  I know that's why I've frequently held off for a
> while in mentioning my problems when others see them - hoping for someone
> here to provide some feedback on what's going on.
> 
> I don't doubt everyone that knows about these things reads the mailing list
> periodically (there's not that much volume!) but when it seems that things
> just aren't heard, it's not all that confidence-inspiring to me, at least.
> Something like "we've seen mention of this ... on LKML that may be causing
> it" would at least give other interested parties here something to look at.
> 
> > Anyways, hang in there and we'll get an answer on these issues sometime in
> > the near future.  Matt is staying on the aacraid and kernel issues rather
> > well for a man who is a proud new pop!!!
> 
> Thanks, good to know that somebody will hopefully provide some feedback
> soon.  Congratulations Matt, I know from experience with my staff here that
> having newborns can be lots of fun - especially juggling them with work!
> 
> Chris
> 
> _______________________________________________
> Linux-PowerEdge mailing list
> Linux-PowerEdge at dell.com
> http://lists.us.dell.com/mailman/listinfo/linux-poweredge
> _______________________________________________
> Linux-aacraid-devel mailing list
> Linux-aacraid-devel at dell.com
> http://lists.us.dell.com/mailman/listinfo/linux-aacraid-devel



More information about the Linux-PowerEdge mailing list