I've been waiting for optimizations like this to get forced along, as they seem to be doing... but eventually I'd like to start using a seeder cache for the distros which we support. No point in doing that early though... it may have concealed some major inefficiencies that have been exposed otherwise. What does everyone else think about the seeder cache idea?

Jeremy

At 12:27 PM 5/10/2004, Jeff Squyres wrote:
See the mails Dave and I exchanged a few minutes ago...

PM/DM was written with the assumption that update-rpms could not handle
multiple requests in a single invocation.  This is an incorrect
assumption.

I'll fix this, but it'll take me a day or three to do so (got some
deadlines today).  It should work fine (aside from the filter issue),
albiet slow.


On Mon, 10 May 2004, Neil Gorsuch wrote:


> I must be misunderstanding what was fixed. On a fresh RH9 vmware box
> with a fresh cvs update -d -P for the oscar tree, update-rpms is still
> showing up using many many minutes of time multiple times. What is being
> cached, or is it being cached at all? How can I tell?
>
> On Fri, 2004-05-07 at 16:23, Lombard, David N wrote:
>
> > A new update-rpms RPM is now in CVS, between a tweak to the setup script
> > to and some changes to update-rpms, I have eliminated the rebuild-cache
> > problem and also eliminated another 100 seconds (due to the above
> > mentioned tweak).
> >
> > I would note that there are many calls to update-rpms, which should be
> > replaceable with either one per uber-package or one total.  This would
> > save a boatload of time.
> >
> > ACTION REQUIRED: Note, there are two libaio RPMS in the base RH distro
> > that need to be removed in the same way the base distro PVM packages are
> > removed.
> >
> > Removing these RPMS gets me to the following error when trying to
> > install the server packages:
> >
> >   /var/tmp/rpm-tmp.16748: line 3: `switcher-reload': not a valid
> > identifier
> >   error: %post(lam-oscar-module-7.0-2) scriptlet failed, exit status 2
> >   /var/tmp/rpm-tmp.96393: line 3: `switcher-reload': not a valid
> > identifier
> >   error: %post(lam-with-gm-oscar-module-7.0-2) scriptlet failed, exit
> > status 2
> >   $pm->install ($dm->query_required_by ()) failed at ./install_server
> > line 63
> >   Couldn't install the required packages needed for OSCAR at
> > ./install_server line 63
> >   --> Step 3: Failed to properly install OSCAR server; please check the
> > logs
> >
> > I have not yet added the disk independence package until I can get to
> > building the client image and reloading a client.
> >
> > I've also tested the node sort routines in OSCAR 3.0, and will add those
> > once 3.1 is functional again.
>
>

--
{+} Jeff Squyres
{+} [EMAIL PROTECTED]
{+} http://www.lam-mpi.org/


------------------------------------------------------- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson & Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 _______________________________________________ Oscar-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/oscar-devel



------------------------------------------------------- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson & Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 _______________________________________________ Oscar-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/oscar-devel

Reply via email to