[Pkg-puppet-devel] Bug#502193: even between 0.24.5-3 instances

[martin f krafft]

forwarded 502193 http://projects.reductivelabs.com/issues/961
tags 502193 upstream
thanks

Doing some more research, I found

  http://projects.reductivelabs.com/issues/961

and cannot help but notice a lot of CLOSE_WAIT connections on all of the
clients experiencing the problem, all held by the daemon process.

tcp       28      0 130.60.75.74:40368      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:40365      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:40366      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51485      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51488      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51484      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:40367      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:33648      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:33650      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51487      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:33651      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51486      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:51489      82.197.185.122:8140     CLOSE_WAIT  26558/ruby
tcp       28      0 130.60.75.74:40364      82.197.185.122:8140     CLOSE_WAIT  26558/ruby

There are not necessarily any matching entries on the server.
Those CLOSE_WAIT connections stay around forever.

If I run puppetd manually with --onetime, they go away with the
process, which could explain why the problem only appears when the
client daemon runs, not when I invoke it manually.

It seems that the fix which closed the upstream bug didn't actually
fix the issue. This is possibly because the commit was refactored
before it went into 0.24.1, possibly introducing a new bug on the
way.

-- 
 .''`.   martin f. krafft <madduck at d.o>      Related projects:
: :'  :  proud Debian developer               http://debiansystem.info
`. `'`   http://people.debian.org/~madduck    http://vcs-pkg.org
  `-  Debian - when you have better things to do than fixing systems
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature (see http://martin-krafft.net/gpg/)
Url : http://lists.alioth.debian.org/pipermail/pkg-puppet-devel/attachments/20090211/8ef5e4a9/attachment-0001.pgp