Date: Wed, 19 Mar 2003 09:52:32 -0500 From: "John Straiton" <jsmailing@clickcom.com> To: "'Andy Farkas'" <andyf@speednet.com.au> Cc: <freebsd-questions@FreeBSD.ORG> Subject: RE: Question about background FSCK Message-ID: <002401c2ee27$304245a0$1916c60a@win2k.clickcom.com> In-Reply-To: <20030319231247.B12616-100000@hewey.af.speednet.com.au>
next in thread | previous in thread | raw e-mail | index | archive | help
Thanks for the idea. While I'm not against the idea of the disk dying,
this is reproduceable quite reliably. Foreground fsck -y in single user
mode works in about 2 minutes (for the 119GB slice) flawlessly every
time and background fsck always hangs the machine.
Additionally, the machine is about a week old Dell Poweredge 1650. While
we all know new != works, it's less likely than a machine with a hard
drive that's been in there awhile.
Unless there's something radically different about how fsck works in
those two fashions, I'm going to assume the reproducability and the fact
that I'm having similar problems on two totally different machines in
different setups (IDE vs SCSI, P4 vs P3, Dell vs HP) means that a dying
disk is not the problem I'm having.
So I ask the list again: Is there a way to disable the background
checking of disks?
John Straiton
jks@clickcom.com
Clickcom, Inc
704-365-9970x101
> -----Original Message-----
> From: Andy Farkas [mailto:andyf@speednet.com.au]
> Sent: Wednesday, March 19, 2003 8:18 AM
> To: John Straiton
> Cc: freebsd-questions@FreeBSD.ORG
> Subject: Re: Question about background FSCK
>
>
> On Tue, 18 Mar 2003, John Straiton wrote:
>
> > While I appreciate the background fsck's that 5.0 provides,
> it appears
> > that there are problems with writing to a drive that is still under
> > the scrutiny of a fsck (tell me if I'm wrong).
> >
> > 'Fer instance, today I brought up a machine that has a 119GB /home
> > partition and then tried to FTP to it. The FTP got to 32kB and
> > hung.... Attempts to reconnect resulted in connections but the
> > inability to STOR.
> >
> > When I do a top, I can see
> >
> > PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU
> > COMMAND
> > 450 root -4 4 632K 376K bufwai 0:01 0.00%
> 0.00% fsck_ufs
> >
> > So I guess it's still running.
>
> Maybe not - 'bufwai' says that its waiting for something and
> 0.00% indicates thats its not actually running anything on the CPU.
>
> My guess is that the disk is dying and taking a long time to
> comlete IO. Do a checkup on your disk..
>
> > Here's the question:
> > I need a solution so that this machine is immediately
> available when
> > it starts taking connections into inetd. What options have
> I on this
> > problem? Is there an override to the write-deny (and if so,
> what risks
> > inclusive to it) or a way to keep the machine from coming
> up until the
> > fsck is done? (ala 4.X style..)
> >
> > I have a machine at home where the boot drive is 160GB that would
> > benefit from the answer as well. If it has to fsck, I have to
> > currently take it to single-user because if I let it
> background fsck,
> > the damn thing will hang (still process packets through the
> NATd but
> > you can't type at all or login for example) a few moments after the
> > login: prompt shows up.
> >
> > Thanks,
> > John Straiton
> > jks@clickcom.com
> > Clickcom, Inc
> > 704-365-9970x101
> >
>
> --
>
> :{ andyf@speednet.com.au
>
> Andy Farkas
> System Administrator
> Speednet Communications
> http://www.speednet.com.au/
>
>
>
>
>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?002401c2ee27$304245a0$1916c60a>