On 7/29/08, Michael Flaig <[EMAIL PROTECTED]> wrote: > Am Montag, den 28.07.2008, 17:46 +0400 schrieb Sergei Golovan: > > > severity 492725 normal > > thanks > > > > On 7/28/08, Michael Flaig <[EMAIL PROTECTED]> wrote: > > > > > > After installation of the package ejabberd is started. However a > > > ejabberdctl command fails. > > > As I noticed epmd is running but beam is not > > > > This is likely a bug #463538 which is fixed in erlang 1:12.b.3-4 (it > > will go to testing soon). > > > okay, I'll watch for the update > Then purge and reinstall, to see what happens.
If you don't want to wait then you can grab erlang-base and erlang-nox packages from here: http://ftp.de.debian.org/debian/pool/main/e/erlang/ (be sure to fetch version 12.b.3-dfsg-4 for your architecture). > > > > > To make ejabberdctl work I needed to "killall empd" > > > and /etc/init.d/ejabberd start > > > > Any distributed erlang service should attach to an existing epmd and > > not run another epmd instance. If you see two epmd daemons then > > something goes wrong. Though I don't know yet where to search for a > > bug. > > > This is also how I understand it, however what happens if rpc is not > available, how would epmd know where to attach to? > > So is it true that beam controlls interprocess communication for > ejabberd trough rpc? ejabberd would be the empd process then, right? No. The opposite. epmd is Erlang portmapepr daemon. Beam is an Erlang virtual machine (with ejabberd inside). > > So if beam is not started that would be the thing to check why. > And having multiple epmd is a result of the rpc not working and > ejabberdctl starting a new epmd when invoked Multiple epmd is something which must not happen. > > Do you have a lenny box to reproduce the reported issue? > And if you would also have a sid installation verify that the problem > does not exist anymore with erlang 1:12.b.3-4? I never saw the multiple epmd, so let's try to understand what happens in your environment. > > > > Are there any log messages about epmd in /var/log/syslog? > > > Jul 29 00:03:55 nagios epmd: epmd: epmd running - daemon = 1 It's a correct message. After this message epmd exits and no new epmd instance runs. > > I did the installation yet again, same result at the 3rd try > However I didn't kill epmd this time but checked the logs > sasl.log is there > ejabberd.log is not > however: telnet localhost 4369 works All is OK (knowing that erlang 12.b.3-dfsg-1 have a bug and can't start services from inside apt-get). Epmd is a daemon which doesn't exit on ejabberd exit (it might serve other Erlang-based services). > > nagios:/var/log/ejabberd# telnet localhost 4369 > Trying 127.0.0.1... <Skipped...> > Escape character is '^]'. > > Connection closed by foreign host. > --- snap --- > > epmd still running and responding after purging. Not good either. All seems OK to me (I think that after upgrading erlang-base and erlang-nox you'll not have to restart ejabberd after apt-get finishes). Epmd shouldn't stop. -- Sergei Golovan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
