[Sisuite-devel] Install Issue...

2006-04-09 Thread David A. Smith
Seems that Install has a little problem Executed: ./install --verbose systemimager-common systemimager-server Had an Error downloading: perl-AppConfig-1.52-4.noarch.rpm (created a file "Failed mirror=belnet.dl.sourceforge.net Seems like the Error Recovery for finding "Subequent Mir

Re: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Finley, Brian E.
I have a fresh ia64 box I should be able to work with in the next couple of days. Cheers, -Brian -- Brian Elliott Finley Linux Strategist, CIS Desk: 630.252.4742 Mobile: 630.631.6621 -Original Message- From: Bernard Li To: Finley, Brian E.; [EMAIL PROTECTED

RE: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Bernard Li
Title: Re: [Sisuite-devel] RE: e100 slowness with trunk Great stuff Andrea - when I did the kernel config for x86_64, I was just using whatever's default.  Anyways glad that you figured it out.   We need to start stabilizing 3.7.3 for the imminent code freeze for OSCAR 5.0.  Are there any

Re: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Finley, Brian E.
Heh, slob. -- Brian Elliott Finley Linux Strategist, CIS Desk: 630.252.4742 Mobile: 630.631.6621 -Original Message- From: Andrea Righi To: Bernard Li CC: sisuite-devel@lists.sourceforge.net; Finley, Brian E. Sent: Sun Apr 09 16:47:13 2006 Subject: Re: [Sisuite

Re: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Andrea Righi
Bernard, stop it! I found the problem! :-) It was the SLOB allocator (http://lwn.net/Articles/157944/); it's more memory efficient than the default SLAB, but it seems that with this allocator rsync becomes very very slow... so better to use SLAB (in fact in the x86_64 config it is correctly enabl

RE: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Bernard Li
Title: Re: [Sisuite-devel] RE: e100 slowness with trunk I'd rather not downgrade the kernel at this point - hopefully newer releases will fix this problem.   rsync is currently at 2.6.7 (we're using 2.6.6), so after the svn server is back up I will try to see if that helps at all...   Che

RE: [Sisuite-devel] reporting of "kernel" in si_monitor

2006-04-09 Thread Bernard Li
Title: Re: [Sisuite-devel] reporting of "kernel" in si_monitor Yeah, that's what I'm guessing too.  That variable name is probably used by something else...   Thanks,   Bernard From: Andrea Righi [mailto:[EMAIL PROTECTED]Sent: Sun 09/04/2006 02:54To: Bernard LiCc: Sisuite-develSubject:

Re: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Andrea Righi
It's a weird problem... with kernel 2.6.15.7 (the kernel just before 2.6.16) rsync works perfectly (I reach about 6-7MB/s). So the reason of the slowness is something that has changed in the particular 2.6.16 kernel relase (I tried also with 2.6.16.2 and the problem persists). At the moment the ri

Re: [Sisuite-devel] download.systemimager.org down?

2006-04-09 Thread Andrea Righi
Erich, the web server seems down... from download.systemimager.org: [EMAIL PROTECTED]:~$ telnet localhost 80 Trying 127.0.0.1... telnet: Unable to connect to remote host: Connection refused [EMAIL PROTECTED]:~$ ps axuw | grep apache arighi 29895 0.0 0.0 176 52 pts/0R+ 08:26 0:00 g

[Sisuite-devel] download.systemimager.org down?

2006-04-09 Thread Erich Focht
Hi, I'm trying to build trunk but am constantly getting "Connection refused" from download.systemimager.org. Is the site down or am I missing something obvious? Thanks, Erich --- This SF.Net email is sponsored by xPML, a groundbreaking script

Re: [Sisuite-devel] RE: e100 slowness with trunk

2006-04-09 Thread Andrea Righi
Bernard, I did some tests and also with my rtl8139 card I found the same slowness using rsync. I didn't notice that because for my previous tests I always used bittorrent. In fact the strange thing is that other protocols doesn't seem to be affected to that slowness!! Moreover it doesn't seem to b

Re: [Sisuite-devel] reporting of "kernel" in si_monitor

2006-04-09 Thread Andrea Righi
Bernard Li wrote: > It seems that durinig init, kernel is correctly reported (uname -r), > however, during imaging, it seems that the string is truncated to just > "2.6" - any idea why this is happening? > Bernard, this should fix the problem.. anyway I don't understand why?? it seems that the "