On 07/12/2011 12:57 PM, Yasha Karant wrote:
On 07/12/2011 11:44 AM, Alexander Hunt wrote:
On 07/12/2011 11:46 AM, Lamar Owen wrote:
On Tuesday, July 12, 2011 01:23:38 PM you wrote:
On Tue, Jul 12, 2011 at 10:14 AM, Lamar Owen<[email protected]> wrote:
These HOWTOS are maintained by Alan Bartlett and Akemi Yagi, who
frequent this list too.
Now that this was mentioned, we have to update them to accommodate
EL6 ... :-)
Indeed. And from what I've seen you write already, that's a pretty
sizeable update.
There is also this post (linked below) in the forum, that makes building
a newer kernel as easy as cutting and pasting the instructions into the
terminal:
http://scientificlinuxforum.org/index.php?showtopic=153
I built (and am now running) 2.6.38.8-35.el6.x86_64 from F15 using this
method, in about 75 minutes, on a Intel CoreDuo with 3 GB Ram. I have
never attempted any kernel revisions or builds in the past.
The method can also be used to rebuild any package from Fedora or EPEL
into an el6 package. I also tried this as well as I needed to update the
linux-firmware package. It's amazingly easy to follow, which reminds me
I have to go there and give Thanks to the author.
Regards to all
A
I do have a small difference of opinion with the instructions in the
mentioned thread.
The instructions have the switch --without debuginfo . I prefer to
build both the regular and debug rpms. This way, if the regular one
throws unexpected exceptions or other diagnostics, one may be able to
isolate or identify the issue. In some cases, it is buggy beta code
from Fedora; in other cases, it can be due to structural
incompatibilities with some portion of a "stock" RHEL distribution
(such as SL 6). One does not have to install the debug version and
after a successful use (test) period has elapsed, one can rebuild
without debug information.
Yasha Karant
I do agree with that, and considered taking out that switch, but as I
was really more interested in testing the method and its viability
rather than having an issue-free kernel, I left it as is. Today I have
been reworking the config file for x86_64 generic and intend to rebuild
the kernel modified for my machine, and with the "--without debuginfo"
switch removed for dubugging purposes. If that all works well, I will
then do as you suggested and rebuild it again with the debuginfo
excluded (after a suitable test period of course).
Regards
Alex
