[gentoo-amd64] Problems after upgrading to 4GB Ram

[gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Lucking]

I'm mananged to solve most Gentoo problems myself in the past - but this is 
really confusing me, so I'd appreciate any suggestions.

I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory, and 
worked perfectly. I upgraded to 4GB, and have spent the last 24 hours trying 
to get everything to work again!

memtest86plus was running overnight, with no errors, so I assume the memory 
is fine.

The system is very unstable now, and will lock up at random intervals - 
during booting, starting firefox, random points compiling kernels, etc etc
etc.

I removed some memory, so I could compile a kernel with MMU options that I 
thought might help, but to no avail.

Additionally, if I go into the BIOS and enable:
    Advance Chipset Features -> MEM Remap (4G RAM Support)
The following appears at the end of dmesg:

skge eth0: enabling interface
skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0
skge unable to clear error (so ignoring them)
skge eth0: Link is up at 100 Mbps, full duplex, flow control tx and rx

and there is no networking - although ifconfig reports the "correct" 
details. With it disabled, networking works.

Any suggestions/pointers would be appreciated, as I feel like I've tried 
everything!

Cheers
Richard 

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Dieter Ries]

hi,
when i upgraded from 512MB to 1024MB, system got unstable, too. in my case the memory-chips didnt macht for dual channel mode. memtest didn't find any errors, but system crashed every 5 minutes. then i disabled dual channel and it worked, perhaps try that.


cu
Dieter


> -------- Original-Nachricht --------
> Datum: Sat, 10 Jun 2006 23:04:59 +0100
> Von: Richard Lucking <lists@lucking.org>
> An: gentoo-amd64@lists.gentoo.org
> Betreff: [gentoo-amd64] Problems after upgrading to 4GB Ram
> 
> I'm mananged to solve most Gentoo problems myself in the past - but this
> is 
> really confusing me, so I'd appreciate any suggestions.
> 
> I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory, and
> worked perfectly. I upgraded to 4GB, and have spent the last 24 hours 
> trying 
> to get everything to work again!
> 
> memtest86plus was running overnight, with no errors, so I assume the
> memory 
> is fine.
> 
> The system is very unstable now, and will lock up at random intervals - 
> during booting, starting firefox, random points compiling kernels, etc etc
> etc.
> 
> I removed some memory, so I could compile a kernel with MMU options that I
> thought might help, but to no avail.
> 
> Additionally, if I go into the BIOS and enable:
>     Advance Chipset Features -> MEM Remap (4G RAM Support)
> The following appears at the end of dmesg:
> 
> skge eth0: enabling interface
> skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0
> skge unable to clear error (so ignoring them)
> skge eth0: Link is up at 100 Mbps, full duplex, flow control tx and rx
> 
> and there is no networking - although ifconfig reports the "correct" 
> details. With it disabled, networking works.
> 
> Any suggestions/pointers would be appreciated, as I feel like I've tried 
> everything!
> 
> Cheers
> Richard 
> 
> -- 
> gentoo-amd64@gentoo.org mailing list

-- 


"Feel free" – 10 GB Mailbox, 100 FreeSMS/Monat ...
Jetzt GMX TopMail testen: http://www.gmx.net/de/go/topmail
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Lucking]

> when i upgraded from 512MB to 1024MB, system got unstable, too. in my case 
> the memory-chips didnt macht for dual channel mode. memtest didn't find 
> any errors, but system crashed every 5 minutes. then i disabled dual 
> channel and it worked, perhaps try that.
>

Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial. It's 
worked perfectly today with any 2 of them - just not when I fit all 4.

It didn't work *at all* when I first put in 4GB, I had to recompile the 
kernel with IOMMU options - now it works, just not under "load".

Cheers
Richard

>> -------- Original-Nachricht --------
>> Datum: Sat, 10 Jun 2006 23:04:59 +0100
>> Von: Richard Lucking <lists@lucking.org>
>> An: gentoo-amd64@lists.gentoo.org
>> Betreff: [gentoo-amd64] Problems after upgrading to 4GB Ram
>>
>> I'm mananged to solve most Gentoo problems myself in the past - but this
>> is
>> really confusing me, so I'd appreciate any suggestions.
>>
>> I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory, 
>> and
>> worked perfectly. I upgraded to 4GB, and have spent the last 24 hours
>> trying
>> to get everything to work again!
>>
>> memtest86plus was running overnight, with no errors, so I assume the
>> memory
>> is fine.
>>
>> The system is very unstable now, and will lock up at random intervals -
>> during booting, starting firefox, random points compiling kernels, etc 
>> etc
>> etc.
>>
>> I removed some memory, so I could compile a kernel with MMU options that 
>> I
>> thought might help, but to no avail.
>>
>> Additionally, if I go into the BIOS and enable:
>>     Advance Chipset Features -> MEM Remap (4G RAM Support)
>> The following appears at the end of dmesg:
>>
>> skge eth0: enabling interface
>> skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0
>> skge unable to clear error (so ignoring them)
>> skge eth0: Link is up at 100 Mbps, full duplex, flow control tx and rx
>>
>> and there is no networking - although ifconfig reports the "correct"
>> details. With it disabled, networking works.
>>
>> Any suggestions/pointers would be appreciated, as I feel like I've tried
>> everything!
>>
>> Cheers
>> Richard
>>
>> -- 
>> gentoo-amd64@gentoo.org mailing list
>
> -- 
>
>
> "Feel free" – 10 GB Mailbox, 100 FreeSMS/Monat ...
> Jetzt GMX TopMail testen: http://www.gmx.net/de/go/topmail
> -- 
> gentoo-amd64@gentoo.org mailing list
>
> 

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Tony Johnson]

On Sunday 11 June 2006 08:04, Richard Lucking wrote:
> I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory, and
> worked perfectly. I upgraded to 4GB, and have spent the last 24 hours
> trying to get everything to work again!
>
> memtest86plus was running overnight, with no errors, so I assume the memory
> is fine.
>
> The system is very unstable now, and will lock up at random intervals -
> during booting, starting firefox, random points compiling kernels, etc etc
> etc.

I have the same motherboard. Initially I had 1Gb RAM, but when I upgraded to 
2Gb I experienced the same type of problems... not so many lock-ups, but 
compilation of anything seemd to stop at random times with random fatal 
errors. Memtest was fine... everything "seemed" fine.

To fix the problem, I replaced the 2Gb "no-name" RAM with 4Gb GEIL RAM, and 
lo-and-behold, not a single problem since... The other RAM has been relegated 
to a 32bit box, and there is no problems.

So, maybe it is just the quality of the components that is the issue.

- tony
-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: Problems after upgrading to 4GB Ram

[Jani Averbach]

On 2006-06-11 00:35+0100, Richard Lucking wrote:
>
> Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial. 
> It's worked perfectly today with any 2 of them - just not when I fit all 4.
> 
> It didn't work *at all* when I first put in 4GB, I had to recompile the 
> kernel with IOMMU options - now it works, just not under "load".

Try to test your system with this:
http://people.redhat.com/dledford/memtest.html

It's a script which stress your system's memory, IO, DMA, etc.

BR, Jani

-- 
Jani Averbach
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Re: Problems after upgrading to 4GB Ram

[Daemon Xavier]

[-- Attachment #1: Type: text/plain, Size: 882 bytes --]

i have seen somethin in my menconifg, that you have to enable if you have
more than 3g of ram

On 6/10/06, Jani Averbach <jaa@jaa.iki.fi> wrote:
>
> On 2006-06-11 00:35+0100, Richard Lucking wrote:
> >
> > Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial.
> > It's worked perfectly today with any 2 of them - just not when I fit all
> 4.
> >
> > It didn't work *at all* when I first put in 4GB, I had to recompile the
> > kernel with IOMMU options - now it works, just not under "load".
>
> Try to test your system with this:
> http://people.redhat.com/dledford/memtest.html
>
> It's a script which stress your system's memory, IO, DMA, etc.
>
> BR, Jani
>
> --
> Jani Averbach
> --
> gentoo-amd64@gentoo.org mailing list
>
>


-- 
Karma, It's Real!
"No penguins were harmed during the writing, just a bunch of broken windows
to let them escape..."-xtacocorex

[-- Attachment #2: Type: text/html, Size: 1343 bytes --]

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Stanke]

I had some problems too. The solution is to use only one and the same type  
of ram. I don't recommend to use different ram chips from different  
manufacturer.

On Sun, 11 Jun 2006 01:03:19 +0200, Dieter Ries <Clip2@gmx.de> wrote:

>
> hi,
> when i upgraded from 512MB to 1024MB, system got unstable, too. in my  
> case the memory-chips didnt macht for dual channel mode. memtest didn't  
> find any errors, but system crashed every 5 minutes. then i disabled  
> dual channel and it worked, perhaps try that.
>
>
> cu
> Dieter
>
>
>> -------- Original-Nachricht --------
>> Datum: Sat, 10 Jun 2006 23:04:59 +0100
>> Von: Richard Lucking <lists@lucking.org>
>> An: gentoo-amd64@lists.gentoo.org
>> Betreff: [gentoo-amd64] Problems after upgrading to 4GB Ram
>>
>> I'm mananged to solve most Gentoo problems myself in the past - but this
>> is
>> really confusing me, so I'd appreciate any suggestions.
>>
>> I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory,  
>> and
>> worked perfectly. I upgraded to 4GB, and have spent the last 24 hours
>> trying
>> to get everything to work again!
>>
>> memtest86plus was running overnight, with no errors, so I assume the
>> memory
>> is fine.
>>
>> The system is very unstable now, and will lock up at random intervals -
>> during booting, starting firefox, random points compiling kernels, etc  
>> etc
>> etc.
>>
>> I removed some memory, so I could compile a kernel with MMU options  
>> that I
>> thought might help, but to no avail.
>>
>> Additionally, if I go into the BIOS and enable:
>>     Advance Chipset Features -> MEM Remap (4G RAM Support)
>> The following appears at the end of dmesg:
>>
>> skge eth0: enabling interface
>> skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0
>> skge unable to clear error (so ignoring them)
>> skge eth0: Link is up at 100 Mbps, full duplex, flow control tx and rx
>>
>> and there is no networking - although ifconfig reports the "correct"
>> details. With it disabled, networking works.
>>
>> Any suggestions/pointers would be appreciated, as I feel like I've tried
>> everything!
>>
>> Cheers
>> Richard
>>
>> --
>> gentoo-amd64@gentoo.org mailing list
>



-- 
^stanke^ (at) ^gentoo64^ (dot) ^sk^

Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Bernhard Auzinger]

Am Sonntag 11 Juni 2006 01:36 schrieb Tony Johnson:
> On Sunday 11 June 2006 08:04, Richard Lucking wrote:
> > I've got a Gigabyte GA-K8NSC-939 motherboard, which had 2GB of memory,
> > and worked perfectly. I upgraded to 4GB, and have spent the last 24 hours
> > trying to get everything to work again!
> >
> > memtest86plus was running overnight, with no errors, so I assume the
> > memory is fine.
> >
> > The system is very unstable now, and will lock up at random intervals -
> > during booting, starting firefox, random points compiling kernels, etc
> > etc etc.
>
> I have the same motherboard. Initially I had 1Gb RAM, but when I upgraded
> to 2Gb I experienced the same type of problems... not so many lock-ups, but
> compilation of anything seemd to stop at random times with random fatal
> errors. Memtest was fine... everything "seemed" fine.
>
> To fix the problem, I replaced the 2Gb "no-name" RAM with 4Gb GEIL RAM, and
> lo-and-behold, not a single problem since... The other RAM has been
> relegated to a 32bit box, and there is no problems.
>
> So, maybe it is just the quality of the components that is the issue.
>
> - tony

I had similar problems when I was upgrading my memory. Finally it was a known 
issue with the Winchester's memory controller. It can't handle fast timings 
in dual channel mode, when all memory banks are filled. So I switched to a 
slower RAM timing and all worked fine.

rgds

Bernhard
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Freeman]

[-- Attachment #1: Type: text/plain, Size: 1119 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Richard Lucking wrote:
> 
> Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial.
> It's worked perfectly today with any 2 of them - just not when I fit all 4.
> 
> It didn't work *at all* when I first put in 4GB, I had to recompile the
> kernel with IOMMU options - now it works, just not under "load".
> 

I ran into a similar issue when it went from 512M -> 1G.  Same brand
RAM, speeds, etc.  I was slightly overclocking at the time and I ended
up fixing things by going back to spec'd speeds.  Make sure you're not
pushing the envelope - no guarantees when you do that and sometimes a
hardware change can keep that from working like it used to.

Sounds like a motherboard issue of some kind - maybe some kind of timing
setting will fix it.  Don't ask me what half of those settings do, though...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEjBcoG4/rWKZmVWkRAvSZAKC4V2zV3mxSPaBKwQV41eiCF0fZmACfQuQK
Y0hI3br+Rzc7KtvBc4T6sB8=
=ogau
-----END PGP SIGNATURE-----

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/x-pkcs7-signature, Size: 4275 bytes --]

[gentoo-amd64] Re: Problems after upgrading to 4GB Ram

[Duncan]

"Richard Lucking" <lists@lucking.org> posted
01f501c68ce6$91c80790$d800a8c0@qaf.cxm, excerpted below, on  Sun, 11 Jun
2006 00:35:33 +0100:

> Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial. It's 
> worked perfectly today with any 2 of them - just not when I fit all 4.
> 
> It didn't work *at all* when I first put in 4GB, I had to recompile the 
> kernel with IOMMU options - now it works, just not under "load".

That sounds very much like...

I don't remember exactly what it's called and it may be slightly different
in your bios anyway, but...

Under the memory timings options, here, it's only visible if I toggle to
manual and get the individual settings...

There's a setting command length or command size or some such, IIRC, with
what should be only two settings, 1T and 2T.  If you have all memory slots
full, it must be 2T, otherwise 1T works fine and is faster.

It sounds like either automatic isn't getting it right, or you had it set
for 1T previously and haven't reset it for 2T.

If that doesn't do it and you have the option, try declocking your memory
a bit.  I had some generic RAM before my last upgrade that was rated
PC3200, that would randomly lockup at that.  As you, memtest86 was
100% clean, and the lockups were usually under load.  When I first got
the memory at the same time I got the board but from a different supplier),
the BIOS didn't have a memory clock limit option, but a BIOS upgrade added
it.  PC3200 is 200MHz clock DDR, so 400MHz memory.  When I clock-limited it
to 183MHz DDR (366MHz, comparable to PC3000), it was solid as a rock.  No
more instabilities, even with tighter than either memory-default or
BIOS-default timings.

Something else that /might/ be it, altho here the IOMMU and AGPGART
wouldn't work /at/ /all/, as in the thing would lockup after loading the
kernel but before loading init if I had it on and either of the above
enabled, it wasn't a random problem at all... the BIOS AGP fast-write. 
(I've no idea at all if this applies to PCI-E.)  With it disabled, both
the above worked as they should.



-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Nuitari]

>> Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial.
>> It's worked perfectly today with any 2 of them - just not when I fit all 4.
>>
>> It didn't work *at all* when I first put in 4GB, I had to recompile the
>> kernel with IOMMU options - now it works, just not under "load".
>>
>
> I ran into a similar issue when it went from 512M -> 1G.  Same brand
> RAM, speeds, etc.  I was slightly overclocking at the time and I ended
> up fixing things by going back to spec'd speeds.  Make sure you're not
> pushing the envelope - no guarantees when you do that and sometimes a
> hardware change can keep that from working like it used to.
>
> Sounds like a motherboard issue of some kind - maybe some kind of timing
> setting will fix it.  Don't ask me what half of those settings do, though...

When I went from 2x 512mb to 2x 512 + 2x1gb, my MSI motherboard 
automatically went to DDR333 speed instead of DDR400, as specified in its 
manual. Dual-channel is still active.
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Bernhard Auzinger]

Am Sonntag 11 Juni 2006 20:10 schrieb Nuitari:
> >> Thanks for the suggestion. They are 4 matching 1GB sticks from Crucial.
> >> It's worked perfectly today with any 2 of them - just not when I fit all
> >> 4.
> >>
> >> It didn't work *at all* when I first put in 4GB, I had to recompile the
> >> kernel with IOMMU options - now it works, just not under "load".
> >
> > I ran into a similar issue when it went from 512M -> 1G.  Same brand
> > RAM, speeds, etc.  I was slightly overclocking at the time and I ended
> > up fixing things by going back to spec'd speeds.  Make sure you're not
> > pushing the envelope - no guarantees when you do that and sometimes a
> > hardware change can keep that from working like it used to.
> >
> > Sounds like a motherboard issue of some kind - maybe some kind of timing
> > setting will fix it.  Don't ask me what half of those settings do,
> > though...
>
> When I went from 2x 512mb to 2x 512 + 2x1gb, my MSI motherboard
> automatically went to DDR333 speed instead of DDR400, as specified in its
> manual. Dual-channel is still active.

Let me guess, you have a AMD64 CPU with Winchester Core? As I wrote this 
morning, the Winchester's memory controller can't handle fast memory timings, 
if all memory banks are filled. Concerning this matter, I wrote an email to 
amd a few months ago and they sad that the winchester core and its 
predecessors have half-baked memory controllers on the chip. You have to set 
the command rate timing from 1T two 2T (as duncan told before) and your 
memory will work in DDR400 mode and without failure.

If you have a venice core or later versions, you should not recognize this 
behaviour.

The limitations of the winchester's memory controller should be listed 
somewhere in the following paper.
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/25759.pdf

rgds

Bernhard
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Lucking]

----- Original Message ----- 
From: "Bernhard Auzinger" <e0026053@student.tuwien.ac.at>

> Let me guess, you have a AMD64 CPU with Winchester Core? As I wrote this
> morning, the Winchester's memory controller can't handle fast memory 
> timings,
> if all memory banks are filled. Concerning this matter, I wrote an email 
> to
> amd a few months ago and they sad that the winchester core and its
> predecessors have half-baked memory controllers on the chip. You have to 
> set
> the command rate timing from 1T two 2T (as duncan told before) and your
> memory will work in DDR400 mode and without failure.

Thanks for all the suggestions. I went into the BIOS and changed the "1T/2T 
DRAM Timing" option from "Auto" to "2T". Still unable to compile a kernel...

In the "MB Intelligent Tweaker" I changed "DDR Clock / Timing Mode" and "DDR 
Timing" to Manual. This gave DDR Speed 333mhz - still crashes before I 
managed to compile a kernel...

I've tried various combinations of RAM, and the same happened whenever I 
have more than 2GB. Also - I double checked with the Kubuntoo Live DVD, and 
the same thing happened after a while.

Anything I may have overlooked? I've emailed Gigabyte on the off chance that 
throws up something, as I know this isn't a gentoo problem as I first 
thought.

Cheers
Richard 

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Freeman]

[-- Attachment #1: Type: text/plain, Size: 1390 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Richard Lucking wrote:
> Anything I may have overlooked? I've emailed Gigabyte on the off chance
> that throws up something, as I know this isn't a gentoo problem as I
> first thought.
> 

When you tested modules on their own, did you try them in the same slots
you use with everything installed?  Or, did you put one module in slot
0, and then after that worked put the other module in slot 0?  It could
be a physical problem with one of your slots.  Granted, your motherboard
might not let you run one with memory only in slot 1, but if it does see
if that makes a difference.  Seems odd, but computers are hardware, and
if there is some kind of defect in the traces you might get some noise
on only one of the slots.

Just grasping at straws, but I figured ideas won't hurt you at this point.

Did you try running memtest86 for a long period of time, or some tool
designed to burn-in RAM?  If the modules are fine on their own it could
be a bad motherboard.  If it is warranted you might consider returning
it.  Might not hurt to try swapping RAM with a friend if you can manage
first.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEkKB0G4/rWKZmVWkRAkT7AJ9GzSa8wrlnnbcslyouOEUQLVA4WwCfYLaC
gAhWHiw1gBxe4cSOz1VjbT0=
=MyCL
-----END PGP SIGNATURE-----

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/x-pkcs7-signature, Size: 4275 bytes --]

Re: [gentoo-amd64] Problems after upgrading to 4GB Ram

[Richard Lucking]

> When you tested modules on their own, did you try them in the same slots
> you use with everything installed?  Or, did you put one module in slot
> 0, and then after that worked put the other module in slot 0?  It could
> be a physical problem with one of your slots.  Granted, your motherboard
> might not let you run one with memory only in slot 1, but if it does see
> if that makes a difference.  Seems odd, but computers are hardware, and
> if there is some kind of defect in the traces you might get some noise
> on only one of the slots.
>
> Just grasping at straws, but I figured ideas won't hurt you at this point.

I've been grasping at plenty! I tried all combinations - and anything over 
2gb caused problems.

> Did you try running memtest86 for a long period of time, or some tool
> designed to burn-in RAM?  If the modules are fine on their own it could
> be a bad motherboard.  If it is warranted you might consider returning
> it.  Might not hurt to try swapping RAM with a friend if you can manage
> first.

The "solution", after much tweaking of settings, appears to be to turn down 
the "CPU/DDR Clock Ratio" to 2/1.33 - anything higher causes lockups. I've 
also "found" the hidden AGP settings, and the system is running quicker than 
ever, so I'm finally happy!

However if I move the memory hole above 4gb the ethernet driver "fails" 
still, but I an live withouth a few hundred mb for the moment.

With all these lockups I seem to have "broken" portage though, after every 
emerge I get:

>>> Original instance of package unmerged safely.
Traceback (most recent call last):
  File "/usr/bin/emerge", line 3505, in ?
    mydepgraph.merge(pkglist)
  File "/usr/bin/emerge", line 2034, in merge
    retval=portage.doebuild(y,"merge",myroot,self.pkgsettings,edebug,tree="porttree")
  File "/usr/lib/portage/pym/portage.py", line 3037, in doebuild
    vartree=vartree, prev_mtimes=prev_mtimes)
  File "/usr/lib/portage/pym/portage.py", line 3246, in merge
    mydbapi=mydbapi, prev_mtimes=prev_mtimes)
  File "/usr/lib/portage/pym/portage.py", line 6693, in merge
    cleanup=cleanup, mydbapi=mydbapi, prev_mtimes=prev_mtimes)
  File "/usr/lib/portage/pym/portage.py", line 6361, in treewalk
    mylock = portage_locks.lockfile(os.path.join(destroot, 
CONFIG_MEMORY_FILE))
  File "/usr/lib/portage/pym/portage_locks.py", line 128, in lockfile
    lockfilename,myfd,unlinkfile,locking_method = 
lockfile(mypath,wantnewlockfile,unlinkfile)

With the last two lines repeated hundreds of times, then:

  File "/usr/lib/portage/pym/portage_locks.py", line 128, in lockfile
    lockfilename,myfd,unlinkfile,locking_method = 
lockfile(mypath,wantnewlockfile,unlinkfile)
  File "/usr/lib/portage/pym/portage_locks.py", line 64, in lockfile
    if not os.path.exists(os.path.dirname(mypath)):
  File "/usr/lib/python2.4/posixpath.py", line 119, in dirname
    return split(p)[0]
  File "/usr/lib/python2.4/posixpath.py", line 79, in split
    if head and head != '/'*len(head):
RuntimeError: maximum recursion depth exceeded in cmp


The package appears to have been installed correctly. I've deleted the locks 
dir, fscked the partition, and rebuilt/regen-ed everything I can think of... 
anything I may have overlooked?

Cheers
Richard 

-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: Problems after upgrading to 4GB Ram

[Duncan]

"Richard Lucking" <lists@lucking.org> posted
006101c69140$404d2590$d800a8c0@qaf.cxm, excerpted below, on  Fri, 16 Jun
2006 13:27:36 +0100:

> With all these lockups I seem to have "broken" portage though, after every
> emerge I get:
> 
>>>> Original instance of package unmerged safely.
> Traceback (most recent call last):
>   File "/usr/bin/emerge", line 3505, in ?
>     mydepgraph.merge(pkglist)
>   File "/usr/bin/emerge", line 2034, in merge
>     retval=portage.doebuild(y,"merge",myroot,self.pkgsettings,edebug,tree="porttree")

That's likely a bad db entry.  Those are in /var/db/pkg/*, so first thing
to do is to umount that partition (you may have to boot the liveCD or
something if it's the root partition) and do a proper fsck.  Once that's
done, try another merge and see if the problem still occurs.

If the problem still occurs, backup the dir and try the process of
elimination, deleting one category at a time, putting it back (from the
backup) if that doesn't solve the problem and trying the next.  Once you
get a category, try the individual packages, then the individual files. 
Eventually, you'll probably find one or more that are corrupt.  You can
try rebuilding the info in them manually if possible, or simply remerge
that package to replace the info.

If that doesn't help, it's probably time to get the portage devs involved.
With the traceback, they should be able to figure out what's going on and
tell you where to go from there.

Note that as long as your world file is fine and your /etc/make.profile
points to your correct profile, the worst that should happen is that
you'll need to do an emerge --emptytree, to rebuild the database from
scratch.  If you have been using FEATURES=buildpkg, you can use emerge
--emptytree --packageonly, and be done in an hour or so, using the binary
packages you've collected to rebuild your system.  If not, it could take
awhile, recompiling every package.  One more reason I think they should
emphasize just how good that feature can be in a pinch, a bit more than
they do in the handbook.



-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@gentoo.org mailing list