[csw-users] NRPE 2009-06-18 / 2009-05-17 incompatibility

[Chris May]

Hi all,

I have servers running NRPE (the nagios remote executor daemon),
almost all of which use version 2.12/2009-05-17 from opencsw (I also
have some linux boxes also, using 2.12 from the ubuntu repos).

Today I updated a couple of boxes, picking up 2.12/2009-06-18 along
the way. Subsequently I found that any box running 05-17 could not
communicate with one running 06-18, and vice versa. In each case, the
server would report

[ID 261363 daemon.error] Error: Request packet had invalid CRC32.
[ID 900267 daemon.error] Client request was invalid, bailing out...

and the client would report

CHECK_NRPE: Received 0 bytes from daemon.  Check the remote server
logs for error messages.

Boxes running 06-18 could communicate with each other fine, though, as
could all the others

I remember an error similar to this when upgrading from NRPE 1 to NRPE
2, but it doesn't seem as if a minor version upgrade (indeed, no
upgrade at all as far as the app version number (2.12) is concerned)
should have such a severe effect.

If I simply replace the /opt/csw/bin/nrpe binary with the one from
05-17 and restart the service, everything works OK again.

Does anyone else see this behaviour? Any ideas what I can do to track
it down an isolate it? Upgrading everything to the 06-18 version isn't
really an option, as it doesn't work with my linux nodes.

Thanks!

Chris