Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Cobalt RaQ/Qube Systems -- Testing Needed
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 6, 7, 8, 9, 10, 11  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures
View previous topic :: View next topic  
Author Message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Wed Jul 28, 2004 2:35 am    Post subject: Reply with quote

DingbatCA wrote:
Question?
Does any one no if the Cobalt RaQ2 is a 64-bit CPU??


AFAIK, it is indeed a 64bit CPU, however it doesn't seem to like 64 linux...
Back to top
View user's profile Send private message
DingbatCA
Guru
Guru


Joined: 07 Jul 2004
Posts: 384
Location: Portland Or

PostPosted: Wed Jul 28, 2004 6:54 am    Post subject: Reply with quote

Thx Mernisse, I will keep it in 32-bit mode

-All
Ok... I now have every thing running, but not stable. With both 2.4.25 and 2.6.6 kernels. Here is my latest bug… Accessing my cobalt through samba or apache I can hard lock my cobalt; no response on serial or ping. Let’s take a look at samba. Below is my smb.conf, I don’t think that is the problem, but any ideas are welcome. If I beat on my network drive (M:\) I can crash the cobalt, I am not talking about crashing samba, I am talking about a full system halt, with NO log output, that part is the worst. Here are my steps…
Boot up the cobalt, start samba.
Map the network drive in Windows XP Pro (c:\net use m: \\cobalt\public /user:foo password)
Run Sandra (http://www.sisoftware.net/) “file system benchmark” on M drive… Hard lock.

I have tried all my troubleshooting tricks, but with out logs, I am dead in the water. Does any one have and idea or even better, a fix??

[global]
workgroup = TEST
server string = cobalt
log file = /var/log/samba/log.%m

[public]
path = /home/public
read only = No
guest ok = Yes

System layout.
Two HDD’s both IBM’s 7200rpm.
30GB reiserfs & ext2, (/), (/boot), (swap)
120GB reiserfs (/home/public)
128mb ram
Back to top
View user's profile Send private message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Wed Jul 28, 2004 11:35 pm    Post subject: Reply with quote

DingbatCA wrote:


-All
Ok... I now have every thing running, but not stable. With both 2.4.25 and 2.6.6 kernels. Here is my latest bug… Accessing my cobalt through samba or apache I can hard lock my cobalt; no response on serial or ping.


sounds alot like the problem I'm having.

with mine, whenever there's alot of network activity across both interfaces the interfaces BOG way down...

are you seeing the hard locks with 2.6.7 as well?
Back to top
View user's profile Send private message
DingbatCA
Guru
Guru


Joined: 07 Jul 2004
Posts: 384
Location: Portland Or

PostPosted: Wed Jul 28, 2004 11:42 pm    Post subject: Reply with quote

-mernisse
I have not tried the 2.6.7 yet. I will see if i can compile it tonight and and run my test. Just for fun and if you have some spair time, try running sandra's file system bench mark aginst a maped network drive (to your cobalt) and see if its just me or if we have a real bug here.
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Thu Jul 29, 2004 1:52 am    Post subject: Reply with quote

I found Apache 2.0 wanted to misbehave on my Qube 2. Have you tried 1.3 at all? 1.3.31 has worked just fine for me -- I might retry 2.0 to see if the bug still exists.

Essentially, I had the problem that Apache wanted to start up 87 threads -- a bit excessive, basically it choked the machine.

Samba also seemed to want to do the same thing, an update seemed to cure the problem however.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
deephack1982
n00b
n00b


Joined: 06 Apr 2004
Posts: 45
Location: Scotland

PostPosted: Thu Jul 29, 2004 5:50 pm    Post subject: Successful Qube 2 install Reply with quote

Hello all,

Thanks to dingbats help I have a successfull install on a qube 2. Running postfix, ssh, courier imap and some other bits and bobs on it. Been up for days with no problems.

Just wondering, alot of the packages I've been using I've had to install manually as they haven't had mips keywords(ie. emerge <ebuild file>). Who should I inform as to their status on mips so that keywords may be added? As far as I can tell the only reason they don't have keywords is that they haven't been tested and I've found them all to work fine.
Back to top
View user's profile Send private message
DingbatCA
Guru
Guru


Joined: 07 Jul 2004
Posts: 384
Location: Portland Or

PostPosted: Thu Jul 29, 2004 10:31 pm    Post subject: Reply with quote

Still not working... :-(
I have samba 3.0.5 installed and apache 1.3.31 installed. I could not get apache 2.0.50 to install. Samba only had 2 process's running smbd and nmbd... Apache has about 8 process's, but keeps its self in line. I have the same crashing problem on different kernels, 2.4.26, 2.6.6, 2.6.7. while i run the benchmark the system hard locks. I have been watching up to the point of lock and it looks like the problem may be disk controler based... recompilling the kernel with out the via chips set support, running the generic chip set. took out ata support and disk DMA. I will take a performance hit... But any thing is better then what i have currently have.
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Fri Jul 30, 2004 12:40 am    Post subject: Re: Successful Qube 2 install Reply with quote

deephack1982 wrote:
Hello all,

Thanks to dingbats help I have a successfull install on a qube 2. Running postfix, ssh, courier imap and some other bits and bobs on it. Been up for days with no problems.

Just wondering, alot of the packages I've been using I've had to install manually as they haven't had mips keywords(ie. emerge <ebuild file>). Who should I inform as to their status on mips so that keywords may be added? As far as I can tell the only reason they don't have keywords is that they haven't been tested and I've found them all to work fine.


You'll have to add the 'mips' keyword to each of the required ebuilds.
Open the ebuild in a text editor, where it says KEYWORDS=, stick mips somewhere in that line.

Or alternatively, use the following script -- this should keyword everything to ~mips...

Code:
for file in `find /usr/portage -name \*.ebuild`; do
    mv -v $file $file.old;
    sed -e 's/^KEYWORDS="/KEYWORDS="~mips /' < $file.old > $file;
done


then emerge the desired package using ACCEPT_KEYWORDS=~mips emerge ...

As for who to tell... perhaps put it up on bugzilla?
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Fri Jul 30, 2004 7:51 pm    Post subject: Reply with quote

DingbatCA wrote:
-mernisse
I have not tried the 2.6.7 yet. I will see if i can compile it tonight and and run my test. Just for fun and if you have some spair time, try running sandra's file system bench mark aginst a maped network drive (to your cobalt) and see if its just me or if we have a real bug here.


well my emerge -u --world just finished, so I'll emerge in samba for the time being and see if I can crash the system...
I somehow doubt my network problems are related to this, but we'll see...
Back to top
View user's profile Send private message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Sun Aug 01, 2004 12:41 am    Post subject: Reply with quote

DingbatCA wrote:
-mernisse
I have not tried the 2.6.7 yet. I will see if i can compile it tonight and and run my test. Just for fun and if you have some spair time, try running sandra's file system bench mark aginst a maped network drive (to your cobalt) and see if its just me or if we have a real bug here.


I ran the test you suggested, and after about 2-3 minutes of smbd taking up about 80% cpu the system hard locked on me...

I don't know what to look for, as there was no errors, just a hard hang.

I posted my .config for 2.6.7 earlier in the thread, I'm using basically the same one for 2.4.26 (which is what is running right now)

These boxes seem to fall on their knees under high I/O...either network or disk...
Back to top
View user's profile Send private message
DingbatCA
Guru
Guru


Joined: 07 Jul 2004
Posts: 384
Location: Portland Or

PostPosted: Mon Aug 02, 2004 3:49 pm    Post subject: Reply with quote

-mernisse
Thx for running the samba test... it prove's i am not the only one with this issue. I have spent over 20 hours trying to find out what is wrong... Every single time I think i have a fix, it explodes on me... OK, for now i am just going to have to live with this issue.

If any one can find a fix for the hard locking problem please post it.
Back to top
View user's profile Send private message
deephack1982
n00b
n00b


Joined: 06 Apr 2004
Posts: 45
Location: Scotland

PostPosted: Fri Aug 20, 2004 2:12 pm    Post subject: Reply with quote

Hello all,

Been running my Qube 2 for weeks now without any real problems. I'm guessing that the lack of posts mean that alot of other people are having few problems too. I do experience the hard locking problem that some people have described but it happens rarely. Trying to sync more than one IMAP folder at a time in kmail seems to do it so I just don't do that. It's strange as system load doens't seem to trigger it. I've had the box at a load of 5 for most of a day and it hasn't locked. Must be an IO thing.

Anyway thanks to everyone for their help and Kumba for his hard work.
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Sat Aug 21, 2004 1:16 am    Post subject: Reply with quote

I haven't had it hard lock as such, but after running for a week, my box seems to stop talking to its SCSI drives. This unfortunately ends up being a hard reboot, because the machine refuses to do a shutdown.

I'm running Linux 2.4.26 from Linux-MIPS CVS. I notice that the HEAD release is 2.6.8.1, so I might give that a try. Last time I tried 2.6, it was 2.6.6 (I think) and it gave me grief with the SCSI controller.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Tue Aug 24, 2004 11:06 pm    Post subject: Reply with quote

deephack1982 wrote:
Hello all,

Been running my Qube 2 for weeks now without any real problems. I'm guessing that the lack of posts mean that alot of other people are having few problems too. I do experience the hard locking problem that some people have described but it happens rarely. Trying to sync more than one IMAP folder at a time in kmail seems to do it so I just don't do that. It's strange as system load doens't seem to trigger it. I've had the box at a load of 5 for most of a day and it hasn't locked. Must be an IO thing.

Anyway thanks to everyone for their help and Kumba for his hard work.


Unfortunatly, due to the problems I've been having with the dual interfaces with the RaQ2+ I have, I had to re-allocate the machine to a different job than it was originally intended for. Since it won't work as a router it's currently running as a Squid proxy server. This it seems to do happily. Pretty soon I'll be migrating the backup DNS and streaming mp3 services to the box to see how it does.

I wish I knew what was going on with the network on the box, but oh well I guess I'll have to deal with it and try to find another box to play router. Thankfully I have a SparcServer 20MP that I can probably press into service as a router / vpn endpoint .
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Mon Sep 06, 2004 2:19 pm    Post subject: Cobalt kernel patch for Linux 2.6.8.1 released Reply with quote

Hi All,
For those who haven't been watching the colonel-panic.org website (I've been stalking it for the last few weeks) the kernel patches for the Linux 2.6.8.1 kernel have been released.

I've got them down, and I'm going to try booting up into 2.6.8.1, see if my ongoing hiccups with SCSI are solved. You might want to give it a try with your boxes, see if it's fixed up any other hiccups with the systems.

Also of interest, there was a recent email reguarding building 64-bit kernels for the Cobalt servers.... with time, and a bit of luck, perhaps we can expect to see these boxes running in full 64-bit mode.

Update:
Just tried Linux 2.6.8.1, so far, so good. I've had a couple of hiccups, mainly with the urandom and clock services (actually, I think it's just the latter... running 'hwclock --utc' locks the machine hard for some odd reason). I'll see about hammering the box with some network & disk traffic, both to the onboard IDE, and the SCSI controller and see what happens.

On booting, I'm hearing the SCSI disks winding up so that's a promising sign. ;-) (they're 1.6" high clunkers, an IBM DGHS, two Seagate ST118273LC and a Seagate ST13181WC) I shall know more when the network services are up.

Update #2:

Done some further tests to try and create the issues a few people have been having:

Previously I was able to move large quantites of data between the SCSI drives, but I hadn't tried any significant IDE traffic. Catting the hard disk device to /dev/null yeilded:

Code:
qube root # time dd if=/dev/hda of=/dev/null
20028960+0 records in
20028960+0 records out

real    16m21.966s
user    3m29.601s
sys     12m20.885s
qube root #


So heavy IDE traffic hasn't bothered it... the drive is a 10GB IBM DeskStar, so to have it read an entire drive in a little over 15 minutes is quite impressive.

The other test was to ping flood the box. You can try this using:
Code:
# ping -s 64000 -f IP-ADDRESS

(-s 64000 specifies the packet Size -- 64KB. -f tells ping to 'flood ping'.)

I tried this earlier this morning for about half an hour without any lockups.
So for those who have been having difficulty, I'd suggest grabbing the latest source from Linux/MIPS CVS and try the patches on colonel-panic.org.

So it looks, at least from where I sit, that the problem may have been solved.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
mernisse
n00b
n00b


Joined: 26 Jun 2004
Posts: 19

PostPosted: Tue Sep 14, 2004 9:22 am    Post subject: Re: Cobalt kernel patch for Linux 2.6.8.1 released Reply with quote

Redhatter wrote:

So it looks, at least from where I sit, that the problem may have been solved.


Well I must say I'm heartened greatly by this news.

I emerged (after unmasking) mips-sources-2.6.8.1 the other night and diffed the tree against the lastest cobalt patches to find that they're applied by the ebuild...
*sweet*

so after a somewhat lengthy make oldconfig, make all, make modules_install I finally reboot the raq...
nothing....nada...zip...colo won't load the image.
so I boot 2.6.7, tweak some things, remerge the source tree and recompile.
Code:

cthulu root # uname -a
Linux cthulu 2.6.8.1-mipscvs-20040814 #3 Sat Sep 11 07:25:36 EDT 2004 mips Nevada V10.0  FPU V10.0 Cobalt RaQ2 GNU/Linux


ok, so far so good.
now, I bring up both interfaces... one is connected to my core network via 2 cisco 2524 routers which are connected by a t1, the other is directly on the core network. Since my biggest issue has been crashing under heavy network load on both interfaces I go and grab a 100M file from my ISP's speedtest server over the t1 link and I grab a 650M iso off an internal box (100M link).

According to my MRTG monitoring of the interfaces, the T1 was pegged at 1.5M/sec for a few minutes and the 100M port was pegged at 12Mbit/sec for a similar amount of time.

the whole time the raq performed admirably and both files were retrieved just fine.

Just for giggles I tried
Code:

cthulu root # time dd if=/dev/hda4 of=/dev/null
real    10m35.159s
user    1m18.516s
sys     7m1.356s
cthulu root # df -h /dev/hda4
Filesystem            Size  Used Avail Use% Mounted on
/dev/hda4             7.0G  2.8G  4.0G  41% /

so all that's left is to setup the ipv6, vpn, and firewall code that I've updated on my current router and make sure it all works without dying again and I might actually get to use the box for what I bought it for!

Either way, thanks to everyone -- gentoo devs (Kumba!!), the great people on the forum, cobalt/sun, ect for making this possible! Kudos for all the hard work.
Back to top
View user's profile Send private message
JeffW_
Retired Dev
Retired Dev


Joined: 19 Sep 2004
Posts: 80
Location: Fremont CA, USA

PostPosted: Tue Sep 21, 2004 10:17 am    Post subject: Reply with quote

Those of us with RaQ 3's (maybe 4's and some Qubes) can run 2.6.x thanks to one day of complete boredom at work...

Code:
illogical@malp illogical $ cat /proc/cobalt/systype ; uname -a
Pacifica
Linux malp 2.6.8.1-cobalt #2 Sun Sep 19 19:43:10 PDT 2004 i586 AMD-K6(tm) 3D processor AuthenticAMD GNU/Linux
Back to top
View user's profile Send private message
Sanguinary
n00b
n00b


Joined: 17 Oct 2004
Posts: 16

PostPosted: Sun Oct 17, 2004 5:18 pm    Post subject: Necessary steps for compiling mips-sources on Raq 2? Reply with quote

I've been using this thread as a reference to try to get Gentoo running on a Raq 2. I can netboot fine, and get a chrooted system up and ready to compile a kernel, but I run into all sorts of errors at that point.

I did get the mips-sources-2.4.27-r1.ebuild sources to compile a kernel, but only after stripping just about everything, including loadable modules support. My problem there is that, no matter what options I give to CoLo, the kernel thinks its being given a "root=/dev/hda1" option, when I tell CoLo to give it "root=/dev/hda3." I can't figure out that one. Booting that way obviously goes nowhere, as there's nothing but CoLo and the kernel on /dev/hda1.

I'd really like to get a 2.6 kernel running. I've not been able to compile this at all. Initially, when I compile the mips-sources-2.6.8.1-r1.ebuild sources, the compile fails due to a missing "mach-gt64120.h." The file actually exists in multiple locations, so with some fiddling, I get past that point. Then it fails much further in the process when compiling "fix-cobalt" (I forgot to save the error).

What is the black art of this process? It seems to require more than getting the mips-sources, because these do not compile cleanly. Yet, obviously, people who know what they're doing can get this to work. (I am not such a person--I have a Gentoo x86 box running, and did not encounter any real challenges there, so I don't really know how to go about solving my problems here.)

I am trying to get the Raq2 to compile the kernel itself, so I've sort of lost patience with these compile failures that come after a whole long time spent compiling.

Thanks.
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 393
Location: Sigma 957

PostPosted: Tue Oct 19, 2004 12:16 am    Post subject: Re: Necessary steps for compiling mips-sources on Raq 2? Reply with quote

Sanguinary wrote:
I've been using this thread as a reference to try to get Gentoo running on a Raq 2. I can netboot fine, and get a chrooted system up and ready to compile a kernel, but I run into all sorts of errors at that point.

I did get the mips-sources-2.4.27-r1.ebuild sources to compile a kernel, but only after stripping just about everything, including loadable modules support. My problem there is that, no matter what options I give to CoLo, the kernel thinks its being given a "root=/dev/hda1" option, when I tell CoLo to give it "root=/dev/hda3." I can't figure out that one. Booting that way obviously goes nowhere, as there's nothing but CoLo and the kernel on /dev/hda1.

To pass kernel commandline args to Colo so it can pass them onto the kernel, do something like this (the following assumes you are starting from the colo command prompt after a fresh reboot):
Code:
> mount
> load <kernel>
> execute root=<root path> console=<console>

Alternatively, you can specify options in menuconfig to be passed to the kernel. I believe it's located underneath the Kernel Debugging menu.


Quote:
I'd really like to get a 2.6 kernel running. I've not been able to compile this at all. Initially, when I compile the mips-sources-2.6.8.1-r1.ebuild sources, the compile fails due to a missing "mach-gt64120.h." The file actually exists in multiple locations, so with some fiddling, I get past that point. Then it fails much further in the process when compiling "fix-cobalt" (I forgot to save the error).

That's odd, because that error should be fixed. You are using the cobalt profile, right? Make sure /etc/make.profile on the cobalt is symlinked to one of the two locations: /usr/portage/profiles/cobalt-mips-2004.1 OR /usr/portage/profiles/default-linux/mips/cobalt/2004.2.

The latter path is the preferred one, as that's for cascading profiles. Using these sets up the USE flags and variables correctly for cobalt support (Adds cobalt to USE, and sets PROFILE_ARCH to "cobalt"). This is used by mips-sources to know when to apply the cobalt patchset to the kernel tree, and that patchset should fix the missing mach-gt64120.h error (iirc, Peter Horton's patchset supplies a blank file that satisfies gcc's request for it).

As for the fix-cobalt error, it sounds like your referring to fixup-cobalt.c in the arch/mips/* portion of the tree. Any errors relating to this should also be fixed if the cobalt patchset gets applied properly.


Quote:
What is the black art of this process? It seems to require more than getting the mips-sources, because these do not compile cleanly. Yet, obviously, people who know what they're doing can get this to work. (I am not such a person--I have a Gentoo x86 box running, and did not encounter any real challenges there, so I don't really know how to go about solving my problems here.)
I am trying to get the Raq2 to compile the kernel itself, so I've sort of lost patience with these compile failures that come after a whole long time spent compiling.

Thanks.

I'll agree with you on this -- compiling a kernel on the raq2 itself can be painfully slow. I use a cross-compiler setup on my x86 box for those needs, but to actually setup such a cross-compiler using the crossdev package requires a new release which I haven't made yet.


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
Sanguinary
n00b
n00b


Joined: 17 Oct 2004
Posts: 16

PostPosted: Tue Oct 19, 2004 11:07 pm    Post subject: Re: Necessary steps for compiling mips-sources on Raq 2? Reply with quote

Kumba wrote:
That's odd, because that error should be fixed. You are using the cobalt profile, right? Make sure /etc/make.profile on the cobalt is symlinked to one of the two locations: /usr/portage/profiles/cobalt-mips-2004.1 OR /usr/portage/profiles/default-linux/mips/cobalt/2004.2.


/etc/make.profile was symlinked to the former, but that didn't seem to be doing the job. When I fixed things, emerging the mips-sources then downloaded a set of cobalt patches, which got things rolling.

I now have a self-booting Raq2 (compiling a default string for the kernel resolved whatever the issue was with CoLo not wanting to pass on my command line options).

Unfortunately, I left out devfs support in my config, so the Gentoo boot scripts give me a stern talking-to. Off to compile another kernel.

Thanks very much for your help.
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 393
Location: Sigma 957

PostPosted: Tue Oct 19, 2004 11:18 pm    Post subject: Re: Necessary steps for compiling mips-sources on Raq 2? Reply with quote

Sanguinary wrote:
Unfortunately, I left out devfs support in my config, so the Gentoo boot scripts give me a stern talking-to. Off to compile another kernel.

Thanks very much for your help.


You can pass "gentoo=nodevfs" to the kernel command line, which sends this to the startup scripts to enable bootup to proceed w/o devfs.

You might also seriously consider udev over devfs, as it is a superior dev management system (and devfs is obseleted anyways, no telling when it'll vanish). See the Gentoo UDev guide on the main site under 'User Docs' for more info on converting. I use udev on my RaQ2 currently, and rather than have to deal with a ~500 item /dev folder, I only have about ~44 under udev.


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
Sanguinary
n00b
n00b


Joined: 17 Oct 2004
Posts: 16

PostPosted: Sat Oct 23, 2004 1:48 am    Post subject: Re: Necessary steps for compiling mips-sources on Raq 2? Reply with quote

Kumba wrote:
You can pass "gentoo=nodevfs" to the kernel command line, which sends this to the startup scripts to enable bootup to proceed w/o devfs.

You might also seriously consider udev over devfs, as it is a superior dev management system (and devfs is obseleted anyways, no telling when it'll vanish). See the Gentoo UDev guide on the main site under 'User Docs' for more info on converting. I use udev on my RaQ2 currently, and rather than have to deal with a ~500 item /dev folder, I only have about ~44 under udev.


Your link turned out to be very familiar to me. I had found it when working on getting my x86 Gentoo box working. I knew there was a reason why I had done what I did initially, compiling the Raq2 kernel without devfs.

No big deal one way or another, I guess. What's another few (or many) hours spent compiling a kernel when I didn't have to?

Thanks for your help on getting this up and running (not just your posts answering my questions, but the work underlying this whole thread).
Back to top
View user's profile Send private message
deephack1982
n00b
n00b


Joined: 06 Apr 2004
Posts: 45
Location: Scotland

PostPosted: Tue Nov 30, 2004 8:51 pm    Post subject: Glibc missing on a Qube 2 Reply with quote

Hello all,

Wondering if anyone can help me here. Almost all my software now produces the following error message.

/lib/librt.so.1: version `GLIBC_2.0' not found

I changed to the /lib directory and ran echo * (my ls don't work no more) and can see librt.so.1 quite happily. I tried running ldconfig and env-update but neither fixed the problem. ldconfig ran without error and env-update did this

root@qube lib # env-update
>>> Regenerating /etc/ld.so.cache...
* Caching service dependencies ...
ls: /lib/librt.so.1: version `GLIBC_2.0' not found (required by ls)
* No scripts to process!
bash: /var/lib/init.d/depcache: No such file or directory
* Failed to cache service dependencies

Any ideas? My poor cube has been working fine for ages now as well.

Thanks for reading!
Back to top
View user's profile Send private message
marmalade
n00b
n00b


Joined: 03 Dec 2004
Posts: 2
Location: Munich - Germany

PostPosted: Fri Dec 03, 2004 10:44 am    Post subject: intalling qube 1 Reply with quote

hi all,
has someone a working kernel for a qube 1 (no serial). I installed a debian kernel image from here
but the netwok card dosn't work!
kumba's kernel dir has only config file, no binaries and i can't get crossdev to work for compiling my own kernel.

thanks!
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Sun Dec 05, 2004 12:59 am    Post subject: Reply with quote

deephack1982 wrote:
Wondering if anyone can help me here. Almost all my software now produces the following error message.

/lib/librt.so.1: version `GLIBC_2.0' not found


Probably your only hope would be to netboot, then perhaps unpack a package for glibc into the root of the drive. As to where to get that package, I might have something here, otherwise, no idea.

The buildpkg feature is very useful for such recoveries :-)

Also, I'd suggest running USE=static emerge busybox when you get it back up, then you can use the Busybox shell if things aren't going well.

marmalade wrote:
has someone a working kernel for a qube 1 (no serial). I installed a debian kernel image from here
but the netwok card dosn't work!
kumba's kernel dir has only config file, no binaries and i can't get crossdev to work for compiling my own kernel.


Are you using patched kernel sources? IIRC last time I tried sticking 2.4.20 on my Qube 2 without patching the kernel, I got a corrupted serial console with no network running.

Also, what is the exact problem you're getting with building a kernel using crossdev? Just saying you can't get it to work isn't exactly helpful.

If glibc is your problem, then you can just make a kernel only cross compiller by using the command:

Code:
# crossdev --arch=mipsel -k


That should build a working cross compiller.
As to building the kernel, I've already started documenting some of this voodoo... My kernel compile guide is here

I've also got a little script that does this for you here.

As always, if you strike troubles... we're here to help :-)
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures All times are GMT
Goto page Previous  1, 2, 3 ... 6, 7, 8, 9, 10, 11  Next
Page 7 of 11

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum