[sage-combinat-devel] Re: combinat out again?

2012-12-04 Thread Keshav Kini
Anne Schilling  writes:
> Hi William,
>
> Since the machine has been down quite a bit and we bought it from the
> NSF grant, do you think it would be a good idea to ask for an exchange?

William sent a mail to the sagemath-users list saying that Dell will be
replacing a faulty (?) memory module in the combinat.math.washington.edu
machine, after which "this will stop happening"::

> It turns out that indeed combinat.math.washington.edu was down with a
> memory error again.  I've rebooted it.Dell will also be fixing it
> (by replacing the DIMM) this week, so this will stop happening.  This
> means some downtime (for combinat only) later this week (not yet
> scheduled).

-Keshav

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



[sage-combinat-devel] Re: combinat out again?

2012-12-13 Thread Dima Pasechnik
On 2012-12-13, William Stein  wrote:
> On Wed, Dec 12, 2012 at 11:30 PM, William Stein  wrote:
>> On Wed, Dec 12, 2012 at 7:00 PM, Anne Schilling  
>> wrote:
>>> Hi William,
>>>
>>> Is combinat.math.washington.edu out once again? The machine does not
>>> seem to respond.
>>
>> It's not crashed due to a memory error, since it responds to ping
>> requests.  However, I can't ssh into it, which can happen when too
>> many people run jobs at once (and the vm.overcommit ratio is too big,
>> and there isn't enough swap).   I'll get the sysadmins to reboot the
>> machine tomorrow morning, then tighten up the vm.overcommit, and add
>> more swap.
>
> The machine is back up.
> The vm.overcommit stuff looks fine -- it's the same settings as on
> sage.math, etc.
>
> Later today, I'm going to add an extra 96GB of swap to the little 32GB
> of swap currently there; this should help with stability a lot.

it might be GAP to blame, in part, as GAP reserves an amount of swap
sort of proportional to available RAM. Now imagine you run 50 Sage
sessions, each with an instance of GAP reserving a chunk of swap.
I tried to convince Volker on #13211 (comments 186 and later) 
that this should be fixed, but he didn't agree...

Dima

>
> William
>
>>
>>  -- William
>>
>>> I saw that several people were running heavy computations on it for the
>>> last week and until yesterday, everything seemed fine.
>>
>>
>>
>>>
>>> Thanks,
>>>
>>> Anne
>>>
>>> On 12/6/12 10:36 AM, William Stein wrote:
 Hi,

 After moving memory around, memtest86 (and the BIOS memtest) detected
 no errors.  If people can try to stress test
 combinat.math.washington.edu for the next 24 hours (especially with
 large-memory computations), that would be very useful!

 William

 On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
> Hi,
>
> Andrew Ohana and I swapped two chips around and are currently running
> memtest86 on
> combinat.math.  I'll check on the results tomorrow at about noon.
>
>  -- William
>
> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>  wrote:
>> Yes, thank you for taking care of this William. I am sure you have better
>> things to do!
>>
>> Andrew
>>
>>
>>
>>
>> --
>> William Stein
>> Professor of Mathematics
>> University of Washington
>> http://wstein.org
>
>
>
> -- 
> William Stein
> Professor of Mathematics
> University of Washington
> http://wstein.org
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



[sage-combinat-devel] Re: combinat out again?

2012-12-13 Thread Volker Braun
On Thursday, December 13, 2012 10:06:47 PM UTC, Dima Pasechnik wrote:

> it might be GAP to blame, in part, as GAP reserves an amount of swap 
> sort of proportional to available RAM. 


Right now it basically eats all available swap, this is going to be fixed 
in #13211
 

> Now imagine you run 50 Sage 
> sessions, each with an instance of GAP reserving a chunk of swap. 


Then the GAP pool will become smaller for each subsequent session as 1/10th 
of the remaining swap (the default choice) gets less and less.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/-CxzWgqATp8J.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-04 Thread William Stein
On Tue, Dec 4, 2012 at 8:39 PM, Keshav Kini  wrote:
> Anne Schilling  writes:
>> Hi William,
>>
>> Since the machine has been down quite a bit and we bought it from the
>> NSF grant, do you think it would be a good idea to ask for an exchange?
>
> William sent a mail to the sagemath-users list saying that Dell will be
> replacing a faulty (?) memory module in the combinat.math.washington.edu
> machine, after which "this will stop happening"::

Yes.  However, they want me to do some more testing first -- they will
either (1) replace a memory chip, or (2) replace the entire
motherboard.  This will be completely covered under the warranty, and
should be done pretty quickly (and certainly within the next 10 days
at worst).

>
>> It turns out that indeed combinat.math.washington.edu was down with a
>> memory error again.  I've rebooted it.Dell will also be fixing it
>> (by replacing the DIMM) this week, so this will stop happening.  This
>> means some downtime (for combinat only) later this week (not yet
>> scheduled).
>
> -Keshav
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sage-combinat-devel" group.
> To post to this group, send email to sage-combinat-devel@googlegroups.com.
> To unsubscribe from this group, send email to 
> sage-combinat-devel+unsubscr...@googlegroups.com.
> For more options, visit this group at 
> http://groups.google.com/group/sage-combinat-devel?hl=en.
>



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-04 Thread Anne Schilling
On 12/4/12 10:42 PM, William Stein wrote:
> On Tue, Dec 4, 2012 at 8:39 PM, Keshav Kini  wrote:
>> Anne Schilling  writes:
>>> Hi William,
>>>
>>> Since the machine has been down quite a bit and we bought it from the
>>> NSF grant, do you think it would be a good idea to ask for an exchange?
>>
>> William sent a mail to the sagemath-users list saying that Dell will be
>> replacing a faulty (?) memory module in the combinat.math.washington.edu
>> machine, after which "this will stop happening"::
> 
> Yes.  However, they want me to do some more testing first -- they will
> either (1) replace a memory chip, or (2) replace the entire
> motherboard.  This will be completely covered under the warranty, and
> should be done pretty quickly (and certainly within the next 10 days
> at worst).

Ok, thanks for letting us know!

Anne

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-04 Thread Andrew Mathas
Yes, thank you for taking care of this William. I am sure you have better 
things to do!

Andrew

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/K5CWg3EqOQ0J.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread William Stein
Hi,

After moving memory around, memtest86 (and the BIOS memtest) detected
no errors.  If people can try to stress test
combinat.math.washington.edu for the next 24 hours (especially with
large-memory computations), that would be very useful!

William

On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
> Hi,
>
> Andrew Ohana and I swapped two chips around and are currently running
> memtest86 on
> combinat.math.  I'll check on the results tomorrow at about noon.
>
>  -- William
>
> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>  wrote:
>> Yes, thank you for taking care of this William. I am sure you have better
>> things to do!
>>
>> Andrew
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "sage-combinat-devel" group.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msg/sage-combinat-devel/-/K5CWg3EqOQ0J.
>>
>> To post to this group, send email to sage-combinat-devel@googlegroups.com.
>> To unsubscribe from this group, send email to
>> sage-combinat-devel+unsubscr...@googlegroups.com.
>> For more options, visit this group at
>> http://groups.google.com/group/sage-combinat-devel?hl=en.
>
>
>
> --
> William Stein
> Professor of Mathematics
> University of Washington
> http://wstein.org



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread Anne Schilling
Hi William,

Sorry to bother you with this, but could you please reset my password
on combinat.math.washington.edu and send it to me? I seem to have forgotten
my old password.

Thank you,

Anne

On 12/6/12 10:36 AM, William Stein wrote:
> Hi,
> 
> After moving memory around, memtest86 (and the BIOS memtest) detected
> no errors.  If people can try to stress test
> combinat.math.washington.edu for the next 24 hours (especially with
> large-memory computations), that would be very useful!
> 
> William
> 
> On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
>> Hi,
>>
>> Andrew Ohana and I swapped two chips around and are currently running
>> memtest86 on
>> combinat.math.  I'll check on the results tomorrow at about noon.
>>
>>  -- William
>>
>> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>>  wrote:
>>> Yes, thank you for taking care of this William. I am sure you have better
>>> things to do!
>>>
>>> Andrew
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "sage-combinat-devel" group.
>>> To view this discussion on the web visit
>>> https://groups.google.com/d/msg/sage-combinat-devel/-/K5CWg3EqOQ0J.
>>>
>>> To post to this group, send email to sage-combinat-devel@googlegroups.com.
>>> To unsubscribe from this group, send email to
>>> sage-combinat-devel+unsubscr...@googlegroups.com.
>>> For more options, visit this group at
>>> http://groups.google.com/group/sage-combinat-devel?hl=en.
>>
>>
>>
>> --
>> William Stein
>> Professor of Mathematics
>> University of Washington
>> http://wstein.org
> 

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread Alex Ghitza
Hi,

On Fri, Dec 7, 2012 at 5:36 AM, William Stein  wrote:
> After moving memory around, memtest86 (and the BIOS memtest) detected
> no errors.  If people can try to stress test
> combinat.math.washington.edu for the next 24 hours (especially with
> large-memory computations), that would be very useful!

OK then, I'll run 4 processes of my patented
hog-the-memory-and-crash-the-server Maeda code for a while.  I'll try
to keep an eye on them, but feel free to kill them if needed.

--
Best,
Alex

--
Alex Ghitza -- Lecturer in Mathematics -- The University of Melbourne
http://aghitza.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread Andrew Mathas
Is anyone able to get an account on combinat.math.washington.edu? I could 
certainly run some memory intensive jobs if required.

Andrew

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/9RLf_AKvIyEJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread William Stein
On Thu, Dec 6, 2012 at 3:48 PM, Andrew Mathas
 wrote:
> Is anyone able to get an account on combinat.math.washington.edu? I could
> certainly run some memory intensive jobs if required.

Are they related to Sage combinatorics research?

William

>
>
> Andrew
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-combinat-devel" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/sage-combinat-devel/-/9RLf_AKvIyEJ.
>
> To post to this group, send email to sage-combinat-devel@googlegroups.com.
> To unsubscribe from this group, send email to
> sage-combinat-devel+unsubscr...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/sage-combinat-devel?hl=en.



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread Andrew Mathas


> Are they related to Sage combinatorics research? 
>

Of course, otherwise I wouldn't ask. (I'm computing the graded dimensions 
of simple modules and the graded dimensions of hom-spaces (for some 
reducible modules).)

Andrew 

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/YgHntXOrxdYJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread William Stein
On Thu, Dec 6, 2012 at 4:27 PM, Andrew Mathas
 wrote:
>
>> Are they related to Sage combinatorics research?
>
>
> Of course, otherwise I wouldn't ask. (I'm computing the graded dimensions of
> simple modules and the graded dimensions of hom-spaces (for some reducible
> modules).)

Send me an offlist email with your desired login name.

>
> Andrew
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-combinat-devel" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/sage-combinat-devel/-/YgHntXOrxdYJ.
>
> To post to this group, send email to sage-combinat-devel@googlegroups.com.
> To unsubscribe from this group, send email to
> sage-combinat-devel+unsubscr...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/sage-combinat-devel?hl=en.



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-06 Thread Nicolas M. Thiery
On Thu, Dec 06, 2012 at 03:48:00PM -0800, Andrew Mathas wrote:
>Is anyone able to get an account on combinat.math.washington.edu?

This machine was purchased using the recent Sage-Combinat NSF
Computional Mathematics grant co-PI'ed by Anne, Dan, Gregg, William
and (unofficially for I am an alien) myself. It's indeed meant for
computations in (algebraic) combinatorics and sage-combinat software
development (e.g. running tests). You definitely fit in :-)

>I could certainly run some memory intensive jobs if required.

Have fun :-)
Nicolas
--
Nicolas M. ThiƩry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-12 Thread Anne Schilling
Hi William,

Is combinat.math.washington.edu out once again? The machine does not
seem to respond.

I saw that several people were running heavy computations on it for the
last week and until yesterday, everything seemed fine.

Thanks,

Anne

On 12/6/12 10:36 AM, William Stein wrote:
> Hi,
> 
> After moving memory around, memtest86 (and the BIOS memtest) detected
> no errors.  If people can try to stress test
> combinat.math.washington.edu for the next 24 hours (especially with
> large-memory computations), that would be very useful!
> 
> William
> 
> On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
>> Hi,
>>
>> Andrew Ohana and I swapped two chips around and are currently running
>> memtest86 on
>> combinat.math.  I'll check on the results tomorrow at about noon.
>>
>>  -- William
>>
>> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>>  wrote:
>>> Yes, thank you for taking care of this William. I am sure you have better
>>> things to do!
>>>
>>> Andrew
>>>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-12 Thread William Stein
On Wed, Dec 12, 2012 at 7:00 PM, Anne Schilling  wrote:
> Hi William,
>
> Is combinat.math.washington.edu out once again? The machine does not
> seem to respond.

It's not crashed due to a memory error, since it responds to ping
requests.  However, I can't ssh into it, which can happen when too
many people run jobs at once (and the vm.overcommit ratio is too big,
and there isn't enough swap).   I'll get the sysadmins to reboot the
machine tomorrow morning, then tighten up the vm.overcommit, and add
more swap.

 -- William

> I saw that several people were running heavy computations on it for the
> last week and until yesterday, everything seemed fine.



>
> Thanks,
>
> Anne
>
> On 12/6/12 10:36 AM, William Stein wrote:
>> Hi,
>>
>> After moving memory around, memtest86 (and the BIOS memtest) detected
>> no errors.  If people can try to stress test
>> combinat.math.washington.edu for the next 24 hours (especially with
>> large-memory computations), that would be very useful!
>>
>> William
>>
>> On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
>>> Hi,
>>>
>>> Andrew Ohana and I swapped two chips around and are currently running
>>> memtest86 on
>>> combinat.math.  I'll check on the results tomorrow at about noon.
>>>
>>>  -- William
>>>
>>> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>>>  wrote:
 Yes, thank you for taking care of this William. I am sure you have better
 things to do!

 Andrew




-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-13 Thread William Stein
On Wed, Dec 12, 2012 at 11:30 PM, William Stein  wrote:
> On Wed, Dec 12, 2012 at 7:00 PM, Anne Schilling  wrote:
>> Hi William,
>>
>> Is combinat.math.washington.edu out once again? The machine does not
>> seem to respond.
>
> It's not crashed due to a memory error, since it responds to ping
> requests.  However, I can't ssh into it, which can happen when too
> many people run jobs at once (and the vm.overcommit ratio is too big,
> and there isn't enough swap).   I'll get the sysadmins to reboot the
> machine tomorrow morning, then tighten up the vm.overcommit, and add
> more swap.

The machine is back up.
The vm.overcommit stuff looks fine -- it's the same settings as on
sage.math, etc.

Later today, I'm going to add an extra 96GB of swap to the little 32GB
of swap currently there; this should help with stability a lot.

William

>
>  -- William
>
>> I saw that several people were running heavy computations on it for the
>> last week and until yesterday, everything seemed fine.
>
>
>
>>
>> Thanks,
>>
>> Anne
>>
>> On 12/6/12 10:36 AM, William Stein wrote:
>>> Hi,
>>>
>>> After moving memory around, memtest86 (and the BIOS memtest) detected
>>> no errors.  If people can try to stress test
>>> combinat.math.washington.edu for the next 24 hours (especially with
>>> large-memory computations), that would be very useful!
>>>
>>> William
>>>
>>> On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
 Hi,

 Andrew Ohana and I swapped two chips around and are currently running
 memtest86 on
 combinat.math.  I'll check on the results tomorrow at about noon.

  -- William

 On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
  wrote:
> Yes, thank you for taking care of this William. I am sure you have better
> things to do!
>
> Andrew
>
>
>
>
> --
> William Stein
> Professor of Mathematics
> University of Washington
> http://wstein.org



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.



Re: [sage-combinat-devel] Re: combinat out again?

2012-12-13 Thread William Stein
On Thu, Dec 13, 2012 at 11:55 AM, William Stein  wrote:
> On Wed, Dec 12, 2012 at 11:30 PM, William Stein  wrote:
>> On Wed, Dec 12, 2012 at 7:00 PM, Anne Schilling  
>> wrote:
>>> Hi William,
>>>
>>> Is combinat.math.washington.edu out once again? The machine does not
>>> seem to respond.
>>
>> It's not crashed due to a memory error, since it responds to ping
>> requests.  However, I can't ssh into it, which can happen when too
>> many people run jobs at once (and the vm.overcommit ratio is too big,
>> and there isn't enough swap).   I'll get the sysadmins to reboot the
>> machine tomorrow morning, then tighten up the vm.overcommit, and add
>> more swap.
>
> The machine is back up.
> The vm.overcommit stuff looks fine -- it's the same settings as on
> sage.math, etc.
>
> Later today, I'm going to add an extra 96GB of swap to the little 32GB
> of swap currently there; this should help with stability a lot.

I changed the vm.overcommit settings to:

vm.overcommit_memory=2
vm.overcommit_ratio=60

and I've added the swap, so now there's 134GB of swap:

root@combinat:/etc# free
 total   used   free sharedbuffers cached
Mem: 198068436   20060256  178008180  0 5854844144288
-/+ buffers/cache:   15330484  182737952
Swap:134537208  0  134537208

Let me know if there's any trouble. I can buy and install another
disk so we have
more swap, if people feel that is a good idea (there is definitely
plenty of room in the
grant for this.)

 -- William

>
> William
>
>>
>>  -- William
>>
>>> I saw that several people were running heavy computations on it for the
>>> last week and until yesterday, everything seemed fine.
>>
>>
>>
>>>
>>> Thanks,
>>>
>>> Anne
>>>
>>> On 12/6/12 10:36 AM, William Stein wrote:
 Hi,

 After moving memory around, memtest86 (and the BIOS memtest) detected
 no errors.  If people can try to stress test
 combinat.math.washington.edu for the next 24 hours (especially with
 large-memory computations), that would be very useful!

 William

 On Wed, Dec 5, 2012 at 3:25 PM, William Stein  wrote:
> Hi,
>
> Andrew Ohana and I swapped two chips around and are currently running
> memtest86 on
> combinat.math.  I'll check on the results tomorrow at about noon.
>
>  -- William
>
> On Tue, Dec 4, 2012 at 10:49 PM, Andrew Mathas
>  wrote:
>> Yes, thank you for taking care of this William. I am sure you have better
>> things to do!
>>
>> Andrew
>>
>>
>>
>>
>> --
>> William Stein
>> Professor of Mathematics
>> University of Washington
>> http://wstein.org
>
>
>
> --
> William Stein
> Professor of Mathematics
> University of Washington
> http://wstein.org



-- 
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.