Bug#453120: mptscsi: Massive workload causes FS to be remounted readonly in VMware guest
Am Donnerstag, den 29.11.2007, 09:30 +0100 schrieb Sebastian Dröge:
> tags 453120 - moreinfo
> Am Dienstag, den 27.11.2007, 18:40 +0100 schrieb maximilian attems:
> > tags 453120 moreinfo
> > severity 453120 normal
> > stop
> > On Tue, Nov 27, 2007 at 03:34:31PM +0100, Sebastian Dröge wrote:
> > > Package: linux-2.6
> > > Severity: important
> > hmm your report is missing important info,
> > you don't tell us which debian version against it is?
> > -> http://wiki.debian.org/DebianKernelReportingBugs
> this is with the etch Kernel, i.e. 2.6.18.dfsg.1-13.
> > > Hi,
> > > when running a Debian system as VMware guest system one could get SCSI
> > > timeouts when there is massive workload on the host system. Since some
> > > kernel version these will get the filesystems remounted read-only, which
> > > probably makes sense for real hardware, but doesn't make sense for
> > > emulated hardware. Instead it should just wait a bit longer.
> > >
> > > See
> > > http://www.tuxyturvy.com/blog/index.php?/archives/31-VMware-ESX-and-ext3-journal-aborts.html
> > >
> > > for some details and what needs to be changed in the kernel. Redhat and
> > > SuSE seem to have this change in their kernels already it seems and
> > > without that change it's impossible to run Debian as VMWare guest
> > > reliable.
> > >
> > > Bye
> > as this page only describes upstream changes,
> > did you try newer linux-2.6 images from backports.org 2.6.22
> > or 2.6.23 from trunk snapshots see apt lines on
> > http://wiki.debian.org/DebianKernel
> Thanks, it's fixed at least in 2.6.23-1~experimental.1~snapshot.9819
> Would be nice if the relevant change could be backported to etch-updates
> or similar. :)
> 2.6.23-1~experimental.1~snapshot.9819 has an equivalent change to the
> one at the URL above.
FYI it's changed in upstream commit