On Sat, Jun 11, 2005 at 02:23:40PM +0900, David Leangen wrote:
> I am finally going to swallow my pride and ask for help...

Don't worry. :)


> I've tried re-reading the docs + mailing lists and reinstalling over
> and over again, but I keep having the same problem. I'm sure there
> is something simple that I must be doing wrong.

Well, let's see.


> I think my problem is related to SSL. When I try running the
> following test,
> I get:
> 
> [EMAIL PROTECTED] conf]# openssl s_client -connect localhost:993 -crlf
> CONNECTED(00000003)
> write:errno=104

Is there a noticeable time delay between the CONNECTED and write
messages?


> I think that the following is related...
> 
> In my bincimap-ssl log file, all the entries look like:
> 
>  tcpserver: fatal: unable to bind: address already used
> 
> When I check via ps, I find:
> 
>   root      5828  0.0  0.0     0    0 ?        Z    14:31   0:00
> [tcpserver] <defunct>

This means that tcpserver can not start because some other program is
already using the port it wants, 993, so tcpserver exits and shows as
defunct since supervise has not yet cleaned up after it. As soon as
supervise cleans up, it tries to start tcpserver again, and the
process repeats.


> Whenever I try to kill the process, another one just pops up until
> I do:
> 
>   svc -d /service/bincimaps/

Right, this tells supervise to stop trying to get tcpserver to run.


> *****[1] bincimap.pem
> 
> -----BEGIN RSA PRIVATE KEY-----
> Proc-Type: 4,ENCRYPTED
> DEK-Info: DES-EDE3-CBC,1C23253E9E96C778
> 
> blablabla
> -----END RSA PRIVATE KEY-----
> 
> -----BEGIN CERTIFICATE REQUEST-----
> blablabla
> -----END CERTIFICATE REQUEST-----

This is wrong, you should have the actual certificate there instead
of the certificate request. But the key is good. This problem does
not cause the tcpserver error however, bincimap.pem isn't read until
later, by bincimap-up, so tcpserver should start just fine.


> *****[2] bincimap.conf

Looks fine.


I'd like some more information to be able to help out. Make sure
supervise knows you want tcpserver down by doing

svc -d /service/bincimaps

then find the pid of the supervise watching bincimaps using ps|grep
supervise.bincimaps (the pid is the first number on the line) and run

strace -fFp pid -o supervise.strace.txt

(replace pid with the actual number)

Then, in another terminal, run

svc -u /service/bincimaps

and check that you see the tcpserver problems again. Let a few of
seconds pass, then down the service again:

svc -d /service/bincimaps

Finally, stop the strace (Ctrl-C) and send the txt file to the list.
(It will probably be rather long, that's ok.)


//Peter

Reply via email to