Passenger settings are currently -

PassengerHighPerformance on
PassengerUseGlobalQueue on

PassengerMaxPoolSize 3
PassengerMaxRequests 1000
PassengerPoolIdleTime 1800

I have run an strace (aka dtruss since these are on OS X clients mostly).

I am seeing this in huge blocks (100s) on the nodes having issues -

__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60
__semwait_signal(0x50B, 0x0, 0x1)         = -1 Err#60

They are also scattered throughout other sections of the run, but in 
smaller 4 or 5 quantities.

The clients vary from 10.6 to 10.8 but the ones having issues are 
consistent with this error.  They are also punctuated with blocks like this 
every once in a while -

select(0x7, 0x7FFF62AECE10, 0x7FFF62AECD90, 0x7FFF62AECD10, 
0x7FFF62AECD00)         = 1 0
connect(0x5, 0x7F84D40FA5A0, 0x10)         = -1 Err#56
fcntl(0x5, 0x4, 0x2)         = 0 0
fcntl_nocancel(0x5, 0x3, 0x0)         = 2 0
fcntl_nocancel(0x5, 0x3, 0x0)         = 2 0
madvise(0x105128000, 0x5C000, 0x9)         = 0 0
fcntl(0x5, 0x3, 0x0)         = 2 0
fcntl(0x5, 0x4, 0x6)         = 0 0
write(0x5, "\200p\001\003\001\0", 0x72)         = 114 0
dtrace: error on enabled probe ID 1791 (ID 116: syscall::read:return): out 
of scratch space in action #12 at DIF offset 44

But that looks like dtrace having issues and nothing to do with Puppet.

The disk on the machines check out clean.  DNS is also good and consistent 
with our working nodes.  These machines are imaged from a standard so 
having a one off is typically very rare, we also use Puppet to help keep 
things nice and tidy.  I am not currently seeing network errors on the 
switches.

This one is quite odd, thanks to everyone who has suggested things.  Very 
much appreciated.

Rob



On Tuesday, February 26, 2013 11:32:33 AM UTC-7, r.yeo wrote:
>
> Running the agent in debug mode on a some clients and I am seeing long run 
> times such as this, but only after the "using pson" lines :
>
> Tue Feb 26 10:57:38 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 10:58:53 2013: Info: 
> /Stage[main]/Vpn::Config/File[/opt/cisco/vpn/.anyconnect_global]: Evaluated 
> in 75.03 seconds
>
> Tue Feb 26 10:58:53 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 11:00:08 2013: Info: 
> /Stage[main]/Amadeuspro/File[/etc/puppet/tmp/com.HairerSoft.AmadeusPro.plist]:
>  
> Evaluated in 74.78 seconds
> Tue Feb 26 11:00:08 2013: Info: 
> /Stage[main]/Adobe/File[/etc/puppet/tmp/com.adobe.Reader.plist]: Starting 
> to evaluate the resource
> Tue Feb 26 11:00:08 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 11:01:23 2013: Info: 
> /Stage[main]/Adobe/File[/etc/puppet/tmp/com.adobe.Reader.plist]: Evaluated 
> in 75.00 seconds
>
> Tue Feb 26 11:01:23 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 11:02:38 2013: Info: /Stage[main]/Puppet::Snow/File[/etc/hooks/
> cleanpup.pl]: Evaluated in 75.01 seconds
>
> This catalog took 921.02 to complete :
>
> Tue Feb 26 11:02:38 2013: Notice: Finished catalog run in 916.78 seconds
> Tue Feb 26 11:02:38 2013: Changes:
> Tue Feb 26 11:02:38 2013: Events:
> Tue Feb 26 11:02:38 2013: Resources:
> Tue Feb 26 11:02:38 2013:             Total: 55
> Tue Feb 26 11:02:38 2013:           Skipped: 6
> Tue Feb 26 11:02:38 2013: Time:
> Tue Feb 26 11:02:38 2013:        Filebucket: 0.00
> Tue Feb 26 11:02:38 2013:           Package: 0.00
> Tue Feb 26 11:02:38 2013:    Ssh authorized key: 0.00
> Tue Feb 26 11:02:38 2013:              User: 0.06
> Tue Feb 26 11:02:38 2013:             Group: 0.11
> Tue Feb 26 11:02:38 2013:              Exec: 0.59
> Tue Feb 26 11:02:38 2013:    Config retrieval: 13.28
> Tue Feb 26 11:02:38 2013:          Last run: 1361901758
> Tue Feb 26 11:02:38 2013:              File: 906.98
> Tue Feb 26 11:02:38 2013:             Total: 921.02
> Tue Feb 26 11:02:38 2013: Version:
> Tue Feb 26 11:02:38 2013:            Config: 1361893868
> Tue Feb 26 11:02:38 2013:            Puppet: 3.1.0
>
> Also, this isn't global as I have nodes getting identical complete configs 
> in 7 seconds and run times are just fine after the using pson lines.
>
> Example of a fast node :
>
> Tue Feb 26 11:22:57 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 11:22:58 2013: Info: 
> /Stage[main]/Amadeuspro/File[/etc/puppet/tmp/com.HairerSoft.AmadeusPro.plist]:
>  
> Evaluated in 0.15 seconds
> Tue Feb 26 11:22:58 2013: Info: 
> /Stage[main]/Adobe/File[/etc/puppet/tmp/com.adobe.Reader.plist]: Starting 
> to evaluate the resource
> Tue Feb 26 11:22:58 2013: Debug: file_metadata supports formats: 
> b64_zlib_yaml pson raw yaml; using pson
> Tue Feb 26 11:22:58 2013: Info: 
> /Stage[main]/Adobe/File[/etc/puppet/tmp/com.adobe.Reader.plist]: Evaluated 
> in 0.07 seconds
>
> Puppet master and clients are 3.1 with Dashboard 1.2.22 and also using 
> PuppetDB.
>
> Any thoughts why I'd have trouble and long run times on select nodes when 
> pson is called?
>
> Thanks!
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/puppet-users?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to