zmperditionctl start asking for password

[k7sle]

With much help from Quanah, we were able to find out the problem.

If you are upgrading from 4.5.9 and do not install the proxy, then later decide to install it, the upgrade script will not properly configure the zmproxy to work, hence my problems.

A huge thanks to Quanah and the entire Zimbra team for the support!

Problem Solved!

[zlex]

ok, you solved it but how? same problem, same scenario

thanks for any reply

[quanah]

Originally Posted by zlex

ok, you solved it but how? same problem, same scenario

What exactly is your scenario? zmperditionctl start doesn't work, or nginx doesn't work? It is expected that zmperditionctl won't work, as it was replaced by nginx (zmproxyctl).

--Quanah

[zlex]

we have installed the 4.5.9 Networkedition and worked 20 users for 4 weeks on it 2 LDAP, 2 SMTP and one Mailbox. after 4 weeks we think every is ok and make an upgrate to 5.0.1 ZCS, migrate our 450 Users from our old 4.0 GPL on debian to the 4.5.9 on Ubuntu 6.06 Networkedition. The complete solution is now running on a HA virtual enviroment.
After the useres are migratet, the performance of the SAN is very bad and we don't find the Problem at the moment. So i set up one more physical Mailbox Server and one more physical MTA and upgrate all servers to the 5.0.2 Version and want to enable imapProxy on the physical Mailbox Server. because i want to move the Mailboxes to the Physical machine, but here are the problems:

1. moving mailboxes created under 4.5.9 works fine. All others created under 5.0.1 is not working. the comes an error.

2. Proxy not work. At first i testet it direct after installation and by testing it via Telnet the connection resets after entering password
The i found this article and edit the nginx.conf.in directly with the entry on line 47

memcache_servers mailboxsrv1.mydomain.net:7072 mailboxsrv2.mydomain.net:7072;

and comment out the dynamic build string, because in the nginx.conf there is only one entry with the mailboxsrv2 without port.

After i restart the Proxyserver the connection was not reset but i can always not authenticate. here the outbut from the nginx.log:

-snip-
2008/02/11 09:15:32 [notice] 21070#0: channel myip2:7072 bad for (at least) 60000 msec, reconnecting
2008/02/11 09:15:32 [notice] 21070#0: attempted re-connection to memcached channel myip2:7072 (rc: -2)
2008/02/11 09:15:52 [notice] 21070#0: channel myip1:7072 bad for (at least) 60000 msec, reconnecting
2008/02/11 09:15:52 [notice] 21070#0: attempted re-connection to memcached channel myip1:7072 (rc: -2)
2008/02/11 09:16:02 [notice] 21070#0: no data on myip2:7072, channel is now bad
2008/02/11 09:16:02 [warn] 21070#0: channel myip2:7072 went bad, will reconnect in 60000 msec
2008/02/11 09:16:22 [notice] 21070#0: no data on myip1:7072, channel is now bad
2008/02/11 09:16:22 [warn] 21070#0: channel myip1:7072 went bad, will reconnect in 60000 msec
2008/02/11 09:17:02 [notice] 21070#0: channel myip2:7072 bad for (at least) 60000 msec, reconnecting
2008/02/11 09:17:02 [notice] 21070#0: attempted re-connection to memcached channel myip2:7072 (rc: -2)


2008/02/11 09:20:44 [info] 21072#0: *3591 client 192.168.100.188 connected to 0.0.0.0:110
2008/02/11 09:20:52 [notice] 21070#0: no data on myip1:7072, channel is now bad
2008/02/11 09:20:52 [warn] 21070#0: channel myip1:7072 went bad, will reconnect in 60000 msec
2008/02/11 09:21:09 [notice] 21072#0: no data on myip1:7072, channel is now bad
2008/02/11 09:21:09 [warn] 21072#0: channel myip1:7072 went bad, will reconnect in 60000 msec
2008/02/11 09:21:09 [info] 21072#0: *3591 client login failed: "Account information not available" while in http auth state, client: 192.168.100.188, server: 0.0.0.0:110, login: "alex@testdomain.de"
-snip-

telnet to port 7072 on both mailbox servers are ok.
Please help, because the zimbra support did not answer and our customers are bord.

regards Alex

[quanah]

Originally Posted by zlex

2. Proxy not work. At first i testet it direct after installation and by testing it via Telnet the connection resets after entering password
The i found this article and edit the nginx.conf.in directly with the entry on line 47

You need to look at the zmupgrade.pm perl module. You will find a number of lines in the "sub upgrade500GA" function relating to the proxy package. Execute those manually as the zimbra user.

--Quanah

[zlex]

Hello Quanah,

what exactly must i run, because some commands in the section isInstalled("zimbra-proxy) are running with enviroemnt variables.

if (main::isInstalled("zimbra-proxy")) {
main::runAsZimbra("$ZMPROV mcf zimbraMemcachedBindPort 11211");

my $zimbraReverseProxyMailHostQuery =
"\(\|\(zimbraMailDeliveryAddress=\${USER}\)\(zimbr aMailAlias=\${USER}\)\)";
my $zimbraReverseProxyDomainNameQuery =
"\(\&\(zimbraVirtualIPAddress=\${IPADDR}\)\(object Class=zimbraDomain\)\)";
my $zimbraReverseProxyPortQuery =
'\(\&\(zimbraServiceHostname=\${MAILHOST}\)\(objec tClass=zimbraServer\)\)';

# We have to use a pipe to write out the Query, otherwise ${USER} gets interpreted
open(ZMPROV, "|su - zimbra -c 'zmprov -l'");
print ZMPROV "mcf zimbraReverseProxyMailHostQuery $zimbraReverseProxyMailHostQuery\n";
print ZMPROV "mcf zimbraReverseProxyPortQuery $zimbraReverseProxyPortQuery\n";
print ZMPROV "mcf zimbraReverseProxyDomainNameQuery $zimbraReverseProxyDomainNameQuery\n";
close ZMPROV;

main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyMailHostAttribute zimbraMailHost");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyPop3PortAttribute zimbraPop3BindPort");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyPop3SSLPortAttribute zimbraPop3SSLBindPort");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyImapPortAttribute zimbraImapBindPort");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyImapSSLPortAttribute zimbraImapSSLBindPort");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyDomainNameAttribute zimbraDomainName");
main::runAsZimbra("$ZMPROV mcf zimbraReverseProxyAuthWaitInterval 10s");
}


Thanks

[quanah]

Originally Posted by zlex

Hello Quanah,

what exactly must i run, because some commands in the section isInstalled("zimbra-proxy) are running with enviroemnt variables.

Then environment variables are escaped. I.e., you don't want them interpreted while running the commands. You want what is in the script in that same form in the LDAP server. For the $ZMPROV variable, look near the top of zmupgrade.pm to get its value.

--Quanah

[zlex]

Hello Quanah,

Ok, i have executet the following lines:

ZMPROV mcf zimbraMemcachedBindPort 11211
ZMPROV mcf zimbraReverseProxyMailHostAttribute zimbraMailHost
ZMPROV mcf zimbraReverseProxyPop3PortAttribute zimbraPop3BindPort
ZMPROV mcf zimbraReverseProxyPop3SSLPortAttribute zimbraPop3SSLBindPort
ZMPROV mcf zimbraReverseProxyImapPortAttribute zimbraImapBindPort
ZMPROV mcf zimbraReverseProxyImapSSLPortAttribute zimbraImapSSLBindPort
ZMPROV mcf zimbraReverseProxyDomainNameAttribute zimbraDomainName
ZMPROV mcf zimbraReverseProxyAuthWaitInterval 10s

zmproxyctl restart

and the config wrotes now the proxy server as memcached server with port 11211 and not the Mailbox servers, is this correct?

when I now want to connect, then i get always the error messages "account information not avalible"

2008/02/15 11:43:16 [info] 28148#0: *10 client login failed: "Account information not available" while in http auth state, client: 192.168.100.188, server: 0.0.0.0:110, login: "user@testdomain.de"

what happend now?

-zlex-

[quanah]

Originally Posted by zlex

Hello Quanah,

Ok, i have executet the following lines:
what happend now?

Well, you didn't say you set the search queries:

print ZMPROV "mcf zimbraReverseProxyMailHostQuery $zimbraReverseProxyMailHostQuery\n";
print ZMPROV "mcf zimbraReverseProxyPortQuery $zimbraReverseProxyPortQuery\n";
print ZMPROV "mcf zimbraReverseProxyDomainNameQuery $zimbraReverseProxyDomainNameQuery\n";


etc...

So it can't find the users when they go to log in. Those queries must be set. As I said previously, they don't depend on current environment variables, those variables are escaped so they can appear literally in the database.

You need to add those.

[zlex]

Originally Posted by quanah

Well, you didn't say you set the search queries:

print ZMPROV "mcf zimbraReverseProxyMailHostQuery $zimbraReverseProxyMailHostQuery\n";
print ZMPROV "mcf zimbraReverseProxyPortQuery $zimbraReverseProxyPortQuery\n";
print ZMPROV "mcf zimbraReverseProxyDomainNameQuery $zimbraReverseProxyDomainNameQuery\n";


etc...

So it can't find the users when they go to log in. Those queries must be set. As I said previously, they don't depend on current environment variables, those variables are escaped so they can appear literally in the database.

You need to add those.

Hello Quanah,

Ok, have understand it now

zimbra@lxpop3e:/root$ zmprov gcf zimbraReverseProxyDomainNameQuery
zimbraReverseProxyDomainNameQuery: \(\&\(zimbraVirtualIPAddress=${IPADDR}\)\(objectCl ass=zimbraDomain\)\)
zimbra@lxpop3e:/root$ zmprov gcf zimbraReverseProxyPortQuery
zimbraReverseProxyPortQuery: \(\&\(zimbraServiceHostname=\${MAILHOST}\)\(object Class=zimbraServer\)\)
zimbra@lxpop3e:/root$ zmprov gcf zimbraReverseProxyMailHostQuery
zimbraReverseProxyMailHostQuery: \(\|\(zimbraMailDeliveryAddress=${USER}\)\(zimbraM ailAlias=${USER}\)\)

is that correct now?
But after restart Zimbra same issue. Account information not available.
I do not find the enviroment variables IPADDR and MAILHOST for user zimbra or root on any servers, so how should find the proxy the servers? must i set this directly?
Mailhosts i think are the Mailbox servers but what is the virtual IP Address?

I believe to give up, because i think i am standing on the powerline to my brain.

-ZLEX-