VM temporary object memory is full

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

VM temporary object memory is full

NorbertHartl
I installed gemston 2.4.4.1 in a virtualbox ubuntu image. I loaded seaside30 (via topaz) into it and now I'm trying to connect with gemtools.

After I log in with gemtools and do a find "ConfigurationOfSeaside30" I get a VM temporary object memory full exception. My installation has the default 50M of temp space. I even raised it to 100M without any effect. I don't have a glue what is wrong with it.

Apart from a standard installation I just forced netldi to fit into the port forwarding scenaria by calling it with

startnetldi -g -a gemstone -p 50378:50378

The OS has plenty of free memory. The machine has 2GB memory and shared memory settings are

kernel.shmmax=838860800
kernel.shmall=2097152

And I found nothing special in the log files that indicates a problem.

Any ideas?

Norbert
Reply | Threaded
Open this post in threaded view
|

Re: VM temporary object memory is full

Paul DeBruicker
On 09/15/2011 10:59 AM, Norbert Hartl wrote:
>   and do a find "ConfigurationOfSeaside30"


are you trying to load Seaside?  If so in gemstone you have to wrap it
like this:


   MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [
     Gofer project load: 'Seaside30'.
   ].



Reply | Threaded
Open this post in threaded view
|

Re: VM temporary object memory is full

NorbertHartl

Am 15.09.2011 um 17:14 schrieb Paul DeBruicker:

> On 09/15/2011 10:59 AM, Norbert Hartl wrote:
>> and do a find "ConfigurationOfSeaside30"
>
>
> are you trying to load Seaside?  If so in gemstone you have to wrap it like this:
>
>
> MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [
>   Gofer project load: 'Seaside30'.
> ].

I loaded seaside alrady exactly how are showing it. My problem occurs if I log in from the gemtools to the stone and try to do anything that I get the exception immediately. I can do a find... for Object and then I get a select list but after clicking one entry I get the exception. So either there is really less memory assigned somehow or it is filled up immediately.

Norbert
Reply | Threaded
Open this post in threaded view
|

Re: VM temporary object memory is full

Dale Henrichs
In reply to this post by NorbertHartl
Norbert,

A dump of the memory allocation and stack at the time the out of memory condition occurs is dumped into the gem log ... you should be able to tell which class /instances are filling memory and then we can go from there in diagnosing the problem ...

Dale

----- Original Message -----
| From: "Norbert Hartl" <[hidden email]>
| To: "GemStone Seaside beta discussion" <[hidden email]>
| Sent: Thursday, September 15, 2011 7:59:30 AM
| Subject: [GS/SS Beta] VM temporary object memory is full
|
| I installed gemston 2.4.4.1 in a virtualbox ubuntu image. I loaded
| seaside30 (via topaz) into it and now I'm trying to connect with
| gemtools.
|
| After I log in with gemtools and do a find "ConfigurationOfSeaside30"
| I get a VM temporary object memory full exception. My installation
| has the default 50M of temp space. I even raised it to 100M without
| any effect. I don't have a glue what is wrong with it.
|
| Apart from a standard installation I just forced netldi to fit into
| the port forwarding scenaria by calling it with
|
| startnetldi -g -a gemstone -p 50378:50378
|
| The OS has plenty of free memory. The machine has 2GB memory and
| shared memory settings are
|
| kernel.shmmax=838860800
| kernel.shmall=2097152
|
| And I found nothing special in the log files that indicates a
| problem.
|
| Any ideas?
|
| Norbert
Reply | Threaded
Open this post in threaded view
|

Re: VM temporary object memory is full

NorbertHartl
Right, I forgot that there is a second place with log files. The stats weren't that informative. But from the log I could see that the temp object memory was set to 10000 which is far to less.

The cause for this is that I started to clean up my netldi start/stop script. I had the impression that netldi does not need to read the full defSeaside environment. I just removed it because the shell was complaining. If some of the environment is missing it is searching for system.conf in

/opt/gemstone/product/data/system.conf

instead of the seaside location. This gave the 10000 setting for a gem.

Now I'm back to source the environment and have the shell complaining. But that is because in a fresh linux box /bin/sh points to /bin/dash instead of /bin/bash. Exchanging the symlink fixes this.

Norbert

Am 15.09.2011 um 18:03 schrieb Dale Henrichs:

> Norbert,
>
> A dump of the memory allocation and stack at the time the out of memory condition occurs is dumped into the gem log ... you should be able to tell which class /instances are filling memory and then we can go from there in diagnosing the problem ...
>
> Dale
>
> ----- Original Message -----
> | From: "Norbert Hartl" <[hidden email]>
> | To: "GemStone Seaside beta discussion" <[hidden email]>
> | Sent: Thursday, September 15, 2011 7:59:30 AM
> | Subject: [GS/SS Beta] VM temporary object memory is full
> |
> | I installed gemston 2.4.4.1 in a virtualbox ubuntu image. I loaded
> | seaside30 (via topaz) into it and now I'm trying to connect with
> | gemtools.
> |
> | After I log in with gemtools and do a find "ConfigurationOfSeaside30"
> | I get a VM temporary object memory full exception. My installation
> | has the default 50M of temp space. I even raised it to 100M without
> | any effect. I don't have a glue what is wrong with it.
> |
> | Apart from a standard installation I just forced netldi to fit into
> | the port forwarding scenaria by calling it with
> |
> | startnetldi -g -a gemstone -p 50378:50378
> |
> | The OS has plenty of free memory. The machine has 2GB memory and
> | shared memory settings are
> |
> | kernel.shmmax=838860800
> | kernel.shmall=2097152
> |
> | And I found nothing special in the log files that indicates a
> | problem.
> |
> | Any ideas?
> |
> | Norbert