[nlug] Java heap space

2014-03-13 Thread Howard White
I have an opportunity to resolve. We use the same install process for all of our application servers. We load the op-sys (CentOS 6) using a pxe-boot process and a standardized kickstart file. So wouldn't you know, I have one server out of 10 on which we may not create a java heap space the

Re: [nlug] Java heap space

2014-03-13 Thread Kent Perrier
The first thing that jumps to my mind is the ulimit on memory size for the user that is running the JVM. Kent On Thu, Mar 13, 2014 at 8:56 AM, Howard White hwh...@vcch.com wrote: I have an opportunity to resolve. We use the same install process for all of our application servers. We load

Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 09:01 AM, Kent Perrier wrote: The first thing that jumps to my mind is the ulimit on memory size for the user that is running the JVM. Kent $ ulimit -a core file size (blocks, -c)0 data seg size (kbytes, -d)unlimited scheduling priority (-e)

Re: [nlug] Java heap space

2014-03-13 Thread Wesley Duffee-Braun
What version of CentOS are you running? I found this https://access.redhat.com/site/solutions/409353 that may be related. On Thu, Mar 13, 2014 at 10:08 AM, Howard White hwh...@vcch.com wrote: On 03/13/2014 09:01 AM, Kent Perrier wrote: The first thing that jumps to my mind is the ulimit on

Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 09:15 AM, Wesley Duffee-Braun wrote: What version of CentOS are you running? I found this https://access.redhat.com/site/solutions/409353 that may be related. yeah, that link looks kinda like this case, except... I booted the system first thing this morning (because free -m

Re: [nlug] Java heap space

2014-03-13 Thread Csaba Toth
JVMs in production can consume many gigs of memory. With 32 bit, you certainly won't be able to provide more than 3-4 gigs. Remember, that because of PAE, the 4GB limit only applies to per process, so the processes together can take far more than 4GB or 12 GB. Memory address fragmentation could be

Re: [nlug] Java heap space

2014-03-13 Thread Wesley Duffee-Braun
Looks like, from https://bugzilla.redhat.com/show_bug.cgi?id=790921, that However, if an enormous number of small files (hundreds of thousands or millions) was being mapped , the address space became extremely fragmented as Csaba surmised. On Thu, Mar 13, 2014 at 10:22 AM, Csaba Toth

Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 09:23 AM, Wesley Duffee-Braun wrote: Looks like, from https://bugzilla.redhat.com/show_bug.cgi?id=790921, that However, if an enormous number of small files (hundreds of thousands or millions) was being mapped , the address space became extremely fragmented as Csaba surmised. I

Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 09:22 AM, Csaba Toth wrote: JVMs in production can consume many gigs of memory. With 32 bit, you certainly won't be able to provide more than 3-4 gigs. Remember, that because of PAE, the 4GB limit only applies to per process, so the processes together can take far more than 4GB or

Re: [nlug] Java heap space

2014-03-13 Thread Kent Perrier
If you don't have access to this KB article let me know and I can print it out for you. Looks to be a kernel bug. A quick cut and paste of the workaround section: This issue is resolved in kernel-2.6.32.358.16.1.el6 and later. Please update the kernel package with yum update kernel. Workaround

Re: [nlug] Java heap space

2014-03-13 Thread Andrew Farnsworth
Since shrinking the memory (and restoring it to the current setting) is simplest and quickest, you might try that first. You could also try going up to 16GB. Andy On Thu, Mar 13, 2014 at 10:27 AM, Howard White hwh...@vcch.com wrote: On 03/13/2014 09:23 AM, Wesley Duffee-Braun wrote: Looks

Re: [nlug] Java heap space

2014-03-13 Thread Brian Pitts
On 03/13/2014 08:56 AM, Howard White wrote: I have an opportunity to resolve. We use the same install process for all of our application servers. We load the op-sys (CentOS 6) using a pxe-boot process and a standardized kickstart file. So wouldn't you know, I have one server out of 10 on

[SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 09:44 AM, Kent Perrier wrote: If you don't have access to this KB article let me know and I can print it out for you. Looks to be a kernel bug. A quick cut and paste of the workaround section: This issue is resolved in |kernel-2.6.32.358.16.1.el6| and later. Please update the

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Kent Perrier
This is actually a good use case for using a configuration management tool like puppet. When you built your new server would would have had a high degree of confidence that it would be identical to the other servers in your environment and you wouldn't have seen this issue! Glad you it got you

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Brian Pitts
Puppet is great, for sure. In this specific case, I'm not sure it would have helped. It's still easy to have drift of packages versions in a Puppetized environment if you aren't methodical about rolling out updates. This is especially true for packages like the kernel, which require

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Steven S. Critchfield
While I would possibly cheer lead for Puppet, I don't think I would have specified anything down to the specific kernel version. I would have looked at dependencies for the tools, and specified all of that, but the kernel is just supposed to work. - Original Message - Puppet is great,

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Howard White
On 03/13/2014 02:16 PM, Brian Pitts wrote: Puppet is great, for sure. In this specific case, I'm not sure it would have helped. It's still easy to have drift of packages versions in a Puppetized environment if you aren't methodical about rolling out updates. This is especially true for packages

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Blake Dunlap
puppet + spacewalk is awesome, just saying =) -Blake On Thu, Mar 13, 2014 at 2:23 PM, Howard White hwh...@vcch.com wrote: On 03/13/2014 02:16 PM, Brian Pitts wrote: Puppet is great, for sure. In this specific case, I'm not sure it would have helped. It's still easy to have drift of

[nlug] QCAD anyone?

2014-03-13 Thread David R. Wilson
Hi guys, I have not found my drafting table (yes, I still have one). Is there anybody in the group that has done stuff with QCAD? I could use a starter course. For that matter, I could use someone that does drafting. It will be billable time. Dave 615 963 4993 office 615 579 3177 cell -- --

Re: [SOLVED] Re: [nlug] Java heap space

2014-03-13 Thread Kent Perrier
The viability of configuration management in any environment is highly dependent on the amount of change in your environment. If you are mostly static (like I am) then the investment in time and effort is hard to justify. Now the new job I start in April is highly dynamic and they use both puppet

Re: [nlug] QCAD anyone?

2014-03-13 Thread Justin W Elam
I have briefly used qcad. Let me know if you need help getting it installed and working a test project. However my drafting friends all use AutoCAD for some reason. -- - Justin W Elam E-mail : justin.w.e...@gmail.com ### On Mar 13, 2014 2:38 PM, David R.

Re: [nlug] QCAD anyone?

2014-03-13 Thread Alex Smith (K4RNT)
​Dassault Systemes offers a free 2D drafting program that can handle .dwg files, called DraftSight. The only thing is that you have to register on their site to get the software activated, but it's a one time activation. They don't lock it down like Microsoft or other DRM systems.​ I'm not a

[nlug] Perl Common Error Handling Subroutine

2014-03-13 Thread Paul Boniol
Probably a question for about a Perl 150 class. My google foo is not finding a solution. - I've got a Perl script that will be run through a web browser. - I want any warnings to go to the browser, so I've used CGI::Carp (and called things appropriately). - I have a common subroutine