Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 24 Feb 2002 13:05:17 -0500 (EST)
From:      Robert Watson <rwatson@FreeBSD.org>
To:        "M. Warner Losh" <imp@village.org>
Cc:        gibbs@scsiguy.com, mi@aldan.algebra.com, luigi@FreeBSD.org, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/etc rc.diskless1 rc.diskless2 
Message-ID:  <Pine.NEB.3.96L.1020224130358.84040D-100000@fledge.watson.org>
In-Reply-To: <20020223.231449.61230251.imp@village.org>

next in thread | previous in thread | raw e-mail | index | archive | help
In order to cause the failure, is significant disk activity required?  IF
so, of what sort?  I can easily set up a diskless box to create file
systems of that ratio and spin booting and using the filesystem, if you
can point me a little in the right direction?  Does it require only
reading, or also writing to the filesystem?  I have little if any
FFS-internals clue, but would be happy to give it a spin (both
figuratively and literally).

Robert N M Watson             FreeBSD Core Team, TrustedBSD Project
robert@fledge.watson.org      NAI Labs, Safeport Network Services

On Sat, 23 Feb 2002, M. Warner Losh wrote:

> In message: <200202240512.g1O5CrI74063@aslan.scsiguy.com>
>             "Justin T. Gibbs" <gibbs@scsiguy.com> writes:
> : >: As far as I  remember, last time this was discussed,  nobody was able to
> : >: offer any evidence  of 4:1 actually causing problems --  there were only
> : >: anecdotal accounts, just like this time...
> : >
> : >No.  I have actual evidence.  I can cause hangs very often and they go
> : >away when I change to 8:1
> : 
> : Then why don't we fix the issue instead of papering over this by
> : changing the default?  If you don't have time to look at this
> : personally, please file a PR with information on how to reproduce
> : the hang.
> 
> I don't have enough information to do it reliably.  It would be
> basically "take one of our flashes, boot until it hangs."  I don't
> have a generic way of causing it to happen because it doesn't happen
> with every flash.  I could file a PR that says that it happens, but
> not a reliable way to reproduce it.
> 
> Warner
> 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1020224130358.84040D-100000>