[gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[Nestor Camacho III]

   
I have one of those Sun Java workstations, and figured I would put 
gentoo on it. While I was installing it I would sometimes get the 
following symptoms.

It would just stop compiling. Would not freeze, I was able to do 
anything else, but it would just stop processing what it was currently 
compiling. When I  alt f2 into another window and do a top, I would see 
that the sh process was at 99%, and had been running for a long time. If 
I do a kill -HUP, on the process things would continue... but I can't 
imgaine that his is a fix for the problem.


Is there anyone out there that has had this problem?

The procs are 2 of the following with 2 gigs of ram.


processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 15
model           : 37
model name      : AMD Opteron(tm) Processor 252
stepping        : 1
cpu MHz         : 2592.666
cache size      : 1024 KB
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge 
mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt 
lm 3dnowext 3dnow pni lahf_lm
bogomips        : 5095.42
TLB size        : 1024 4K pages
clflush size    : 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp




top - 09:49:14 up 1 day, 23:07,  7 users,  load average: 1.00, 1.16, 1.36
Tasks: 108 total,   2 running, 105 sleeping,   0 stopped,   1 zombie
Cpu(s): 15.7% us,  0.5% sy,  7.2% ni, 76.2% id,  0.1% wa,  0.0% hi,  0.2% si
Mem:   2056096k total,  1643992k used,   412104k free,   458632k buffers
Swap:   995988k total,      208k used,   995780k free,   441620k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
*21166 root      25   0  4232  428  340 R 99.4  0.0 827:12.92 sh*
21278 jedi      16   0  188m  67m  12m S  2.0  3.3   0:11.35 python
    1 root      16   0  2548  480  412 S  0.0  0.0   0:00.41 init
    2 root      RT   0     0    0    0 S  0.0  0.0   0:00.18 migration/0
    3 root      34  19     0    0    0 S  0.0  0.0   0:00.02 ksoftirqd/0
    4 root      RT   0     0    0    0 S  0.0  0.0   0:00.22 migration/1
    5 root      34  19     0    0    0 S  0.0  0.0   0:00.02 ksoftirqd/1
    6 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 events/0
    7 root      10  -5     0    0    0 S  0.0  0.0   0:00.25 events/1
    8 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 khelper
    9 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kthread
   14 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kacpid
  103 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kblockd/0
  104 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kblockd/1
  107 root      15   0     0    0    0 S  0.0  0.0   0:00.00 khubd
  175 root      20   0     0    0    0 S  0.0  0.0   0:00.00 pdflush

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[Marco Matthies]

Hi, i've never experienced the same problem nor do I have a dual Opteron 
setup, but I'll try to help:

Nestor Camacho III wrote:
> It would just stop compiling. Would not freeze, I was able to do 
> anything else, but it would just stop processing what it was currently 
> compiling. When I  alt f2 into another window and do a top, I would see 
> that the sh process was at 99%, and had been running for a long time. If 
> I do a kill -HUP, on the process things would continue... but I can't 
> imgaine that his is a fix for the problem.

Seems like an infinite loop perhaps, but might as well be something 
else... As always, it might be software or hardware :)

Do these problems reliably show up with a particular ebuild? You should 
be able to tell which ebuild it was by looking at the end of 
/var/log/emerge.log -- if yes, it might be the fault of that ebuild and 
it would be interesting to know if you can reliably cause it to spin 
forever.

If not it would probably have to be kernel and/or hardware (unless 
somehow the ebuild is behaving non-deterministically). Anything special 
noted in the output of dmesg or /var/log/messages that might indicate 
faulty hardware (for example, memory)? Though hardware failure can have 
lots of reasons (memory, psu, bad/dying capacitors). You might perhaps 
want to try a few nights of memtest on it. I also don't know what 
problems bad sectors on a hdd could cause for an ebuild (maybe something 
like trying to create a file and always failing?).

> Is there anyone out there that has had this problem?

I also vaguely remember seeing some bugs on the kernel.org bugzilla that 
were about smp-kernels on opterons having strange random segfaults (due 
to apparently some small error in their tlb). I don't know if the fixes 
have already been folded into the current gentoo and vanilla kernels or 
if your processor is actually affected (btw, which kernel are you 
running?). Maybe someone who is running smp can comment?

>  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> *21166 root      25   0  4232  428  340 R 99.4  0.0 827:12.92 sh*
> 21278 jedi      16   0  188m  67m  12m S  2.0  3.3   0:11.35 python

I'm a bit puzzled by those two lines, did you run emerge as a regular 
user (seems like it or is that python process something else)? I don't 
really know how good running emerge is supposed to work when not being 
run as root.

Good luck,
Marco
-- 
gentoo-amd64@gentoo.org mailing list

RE: [gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[Bob Young]

Nestor Camacho III wrote:
> It would just stop compiling. Would not freeze, I was able to do
> anything else, but it would just stop processing what it was currently
> compiling. When I  alt f2 into another window and do a top, I would see
> that the sh process was at 99%, and had been running for a long time. If
> I do a kill -HUP, on the process things would continue... but I can't
> imgaine that his is a fix for the problem.

I just joined this list about an hour ago, so I haven't seen the orginal
post. Thought I'd pass along the experience I had recently installing Gentoo
on a dual Opteron box, as it seems it may be relevant. The first week was
very frustrating due to the fact that compiles would just stop at various
randomly chosen points during either the stage1 or stage2 build. Compiler
would just stop, no error message, no error indication of any kind, just no
output, though a ctrl-c would return to the command prompt.

Being the stickler for getting a clean build that I am, I would abort the
install and start over with different use flags and/or compile switches, or
various other changes. Nothing resolved the hang untill I disabled SMP and
the IOAPIC by adding "nosmp" and "noapic" to the command line of the install
CD.

I've done several re-installs since then, for various reasons, all related
to deciding I wanted something different. The hang has never happened since.
Note this only seems to ba a problem with the kernel on the install CD, as
after one particular install I sucessfully emerged kde-meta (about 36 hours
of downloading and compiling)  with the SMP enabled kernel I built.

FWIW

Bob Young
San Jose CA




-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[Nestor Camacho III]

Hmm.... this sounds a lot like my problem... I have managed to get my 
box up, but I still get the hangs as mentioned in my post outside of the 
live cd kernel. I am running an smp kernel. Using gentoo 2.6.12.r10. I 
will upgrade to the latest 2.6.13.r3 and see if that cleans things up.

thanks for the info. If anyone else has any more please chime in! :)

Nes++


Bob Young wrote:

>Nestor Camacho III wrote:
>  
>
>>It would just stop compiling. Would not freeze, I was able to do
>>anything else, but it would just stop processing what it was currently
>>compiling. When I  alt f2 into another window and do a top, I would see
>>that the sh process was at 99%, and had been running for a long time. If
>>I do a kill -HUP, on the process things would continue... but I can't
>>imgaine that his is a fix for the problem.
>>    
>>
>
>I just joined this list about an hour ago, so I haven't seen the orginal
>post. Thought I'd pass along the experience I had recently installing Gentoo
>on a dual Opteron box, as it seems it may be relevant. The first week was
>very frustrating due to the fact that compiles would just stop at various
>randomly chosen points during either the stage1 or stage2 build. Compiler
>would just stop, no error message, no error indication of any kind, just no
>output, though a ctrl-c would return to the command prompt.
>
>Being the stickler for getting a clean build that I am, I would abort the
>install and start over with different use flags and/or compile switches, or
>various other changes. Nothing resolved the hang untill I disabled SMP and
>the IOAPIC by adding "nosmp" and "noapic" to the command line of the install
>CD.
>
>I've done several re-installs since then, for various reasons, all related
>to deciding I wanted something different. The hang has never happened since.
>Note this only seems to ba a problem with the kernel on the install CD, as
>after one particular install I sucessfully emerged kde-meta (about 36 hours
>of downloading and compiling)  with the SMP enabled kernel I built.
>
>FWIW
>
>Bob Young
>San Jose CA
>
>
>
>
>  
>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Problems compiling on Sun Java workstation with dual Opterons.

[Craig Webster]

Hi All,

On 20 Oct 2005, at 01:21, Nestor Camacho III wrote:
> Hmm.... this sounds a lot like my problem... I have managed to get  
> my box up, but I still get the hangs as mentioned in my post  
> outside of the live cd kernel. I am running an smp kernel. Using  
> gentoo 2.6.12.r10. I will upgrade to the latest 2.6.13.r3 and see  
> if that cleans things up.
>
> thanks for the info. If anyone else has any more please chime in! :)

I've only just finished installing and am emerging my apps and  
getting these hangs on 2.6.13-r3 as well. I'm running an SMP enabled  
kernel although oddly the HyperThreading CPU isn't showing up (it did  
on the LiveCD).

Anyone got any ideas what I left out? I can't currently post  
my .config but will get that tomorrow if needed.

Yours,
Craig
--
Craig Webster | t: +44 (0)131 516 8595 | e: craig@xeriom.net
Xeriom.NET    | f: +44 (0)709 287 1902 | w: http://xeriom.net


-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: Problems compiling on Sun Java workstation with dual Opterons.

[Duncan]

Craig Webster posted <1256A32E-EDAC-4792-B229-57AF3B2AA9C3@xeriom.net>,
excerpted below,  on Thu, 20 Oct 2005 01:26:18 +0100:

> I've only just finished installing and am emerging my apps and  
> getting these hangs on 2.6.13-r3 as well. I'm running an SMP enabled  
> kernel although oddly the HyperThreading CPU isn't showing up (it did  
> on the LiveCD).
> 
> Anyone got any ideas what I left out? I can't currently post  
> my .config but will get that tomorrow if needed.

Hyperthreading?  AMD64 CPUs don't need nor should they be configured with
that.  If you have an Intel, with it's deeper pipeline, some of /them/ use
hyperthreading.

Assuming it IS an Intel...

You probably enabled SMP (symmetric multi-processing support) but not SMT
(Symmetric multi-threading --- Hyperthreading) scheduler support.  The
latter is a suboption of the former, only available if SMP is enabled.

Those options are under processor type and features under make menuconfig,
the symbol in .config is CONFIG_SCHED_SMT.  Again, it should be disabled
for true AMD64s, possibly enabled for certain Intel CPUs.

-- 
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 in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html


-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Re: Problems compiling on Sun Java workstation with dual Opterons.

[Craig Webster]

Hi Duncan,

Thanks for your reply.

On 20 Oct 2005, at 05:27, Duncan wrote:
> Craig Webster posted <1256A32E-EDAC-4792- 
> B229-57AF3B2AA9C3@xeriom.net>,
> excerpted below,  on Thu, 20 Oct 2005 01:26:18 +0100:
>
>> I've only just finished installing and am emerging my apps and
>> getting these hangs on 2.6.13-r3 as well. I'm running an SMP enabled
>> kernel although oddly the HyperThreading CPU isn't showing up (it did
>> on the LiveCD).
>>
>> Anyone got any ideas what I left out? I can't currently post
>> my .config but will get that tomorrow if needed.
>
> Assuming it IS an Intel...

It is; It's a Xeon.

> You probably enabled SMP (symmetric multi-processing support) but  
> not SMT
> (Symmetric multi-threading --- Hyperthreading) scheduler support.  The
> latter is a suboption of the former, only available if SMP is enabled.

I have both enabled -- kernel config is up at
http://xeriom.net/~craig/kernel-config.txt

Thanks for your help :)

Yours,
Craig
--
Craig Webster | t: +44 (0)131 516 8595 | e: craig@xeriom.net
Xeriom.NET    | f: +44 (0)709 287 1902 | w: http://xeriom.net


-- 
gentoo-amd64@gentoo.org mailing list