I've been affected by this bug. I'm using etch with a backport of boinc
5.10.44-1 and a custom 2.6.24 kernel. It seems that
the /usr/share/boinc-client/udev-cpu_share script is not getting run or not
working on my system, /etc/init.d/boinc-client status reports cpu_share: 1024
I tried sudo sh
Hi Daniel,
On Friday 22 February 2008 00:04, Daniel Hahler wrote:
Frank S. Thomas wrote:
Why is udev-usr_share installed into /usr/lib instead of /usr/share
although it is architecture-independent?
Probably I've seen it on another package or example done this way.
Please move it to
On Sunday 17 February 2008 19:33, James Westby wrote:
I am attaching the patch used in Ubuntu to work around this issue
for your consideration, and so I am also setting the patch tag
again.
The author of the patch, Daniel Hahler, acknowledges that it is a
workaround, but I believe it is a
On 2/18/08, James Westby [EMAIL PROTECTED] wrote:
tags 461630 patch
thanks
Hi,
I am attaching the patch used in Ubuntu to work around this issue
for your consideration, and so I am also setting the patch tag
again.
The author of the patch, Daniel Hahler, acknowledges that it is a
Frank S. Thomas wrote:
I am attaching the patch used in Ubuntu to work around this issue
for your consideration, and so I am also setting the patch tag
again.
The author of the patch, Daniel Hahler, acknowledges that it is a
workaround, but I believe it is a working solution to the issue.
tags 461630 patch
thanks
Hi,
I am attaching the patch used in Ubuntu to work around this issue
for your consideration, and so I am also setting the patch tag
again.
The author of the patch, Daniel Hahler, acknowledges that it is a
workaround, but I believe it is a working solution to the issue.
Just for the record, this bug originated from
https://bugs.launchpad.net/bugs/177713, which has
some additonal information regarding this report.
signature.asc
Description: This is a digitally signed message part.
On 1/21/08, Frank S. Thomas [EMAIL PROTECTED] wrote:
Just for the record, this bug originated from
https://bugs.launchpad.net/bugs/177713, which has
some additonal information regarding this report.
Correct. This is not a boinc bug. This is due to the changed behaviour
of the linux scheduler.
Package: boinc-client
Version: 5.10.30-2
Severity: important
Tags: patch
--- Please enter the report below this line. ---
The CFS in Linux 2.6.24 has the side effect that the boinc user's idle process
takes about half of the cpu cycles.
This causes severe performance regressions when running
9 matches
Mail list logo
