zmcontrol status returns "Cannot determine services - exiting"

[TMH]

I've been searching for a solution to this for the last few hours, and I've seen loads of threads with this error in them, some of them duplicate this problem precisely, but none of them contain advice that can solve my problem.

When I run zmcontrol start, it will only start the ldap daemon:

Host steve.ac
Starting ldap...Done.

These are the only lines that appear in zimbra.log
Jun 10 18:46:13 vps zimbramon[18278]: 18278:info: Starting services
Jun 10 18:46:14 vps slapd[18378]: @(#) $OpenLDAP: slapd 2.3.34 (Apr 4 2007 12:09:30) $ root@support4.liquidsys.com:/home/build/p4/main/ThirdParty/openldap/openldap-2.3.34/servers/slapd
Jun 10 18:46:14 vps slapd[18379]: slapd starting

It's worth noting that I have absolutely no idea what root@support4.liquidsys.com is. It is certainly not a path that exists on my system.


However, I can now stop it:

[zimbra@vps ~]$ zmcontrol stop
Host steve.ac
Stopping mta...Done
Stopping spell...Done
Stopping snmp...Done
Stopping antivirus...Done
Stopping antispam...Done
Stopping imapproxy...Done
Stopping mailbox...Done
Stopping logger...Done
Stopping ldap...Done

An interesting observation: When I installed Zimbra, I did not select spell (because it kept failing with no error information), and I did not select snmp (because I do not need it). I don't know why those two are showing up above.


The log file now has a few more interesting lines added to it:

Jun 10 19:01:13 vps zimbramon[26537]: 26537:info: Stopping services
Jun 10 19:01:13 vps zimbramon[26537]: 26537:info: Stopping mta
Jun 10 19:01:13 vps postfix[27679]: fatal: chdir(/usr/libexec/postfix): No such file or directory
Jun 10 19:01:14 vps zimbramon[26537]: 26537:info: Stopping spell
Jun 10 19:01:15 vps zimbramon[26537]: 26537:info: Stopping snmp
Jun 10 19:01:15 vps zimbramon[26537]: 26537:info: Stopping antivirus
Jun 10 19:01:15 vps zimbramon[26537]: 26537:info: Stopping antispam
Jun 10 19:01:15 vps zimbramon[26537]: 26537:info: Stopping imapproxy
Jun 10 19:01:18 vps zimbramon[26537]: 26537:info: Stopping mailbox
Jun 10 19:01:18 vps zmtomcatmgr[28141]: threaddump requested
Jun 10 19:01:18 vps zmtomcatmgr[28141]: stale pid 3578 found in /opt/zimbra/log/zmtomcatmgr.pid: No such process
Jun 10 19:01:18 vps zmtomcatmgr[28141]: assuming no other instance is running
Jun 10 19:01:18 vps zmtomcatmgr[28141]: no manager process is running
Jun 10 19:01:20 vps zmtomcatmgr[28148]: stop requested
Jun 10 19:01:20 vps zmtomcatmgr[28148]: stale pid 3578 found in /opt/zimbra/log/zmtomcatmgr.pid: No such process
Jun 10 19:01:20 vps zmtomcatmgr[28148]: assuming no other instance is running
Jun 10 19:01:20 vps zmtomcatmgr[28148]: no manager process is running
Jun 10 19:01:20 vps zimbramon[26537]: 26537:info: Stopping logger
Jun 10 19:01:20 vps zimbramon[26537]: 26537:info: Stopping ldap
Jun 10 19:01:20 vps slapd[24078]: daemon: shutdown requested and initiated.
Jun 10 19:01:20 vps slapd[24078]: slapd shutdown: waiting for 0 threads to terminate
Jun 10 19:01:20 vps slapd[24078]: slapd stopped.


This is my hosts file:
127.0.0.1 localhost localhost.localdomain


Zmhostname:
steve.ac


There is an MX line in my /var/named/steve.ac.db (this is not the entirety of the file ).
steve.ac. 14400 IN MX 0 steve.ac.


My resolv.conf file:
nameserver 208.67.222.222
nameserver 208.67.220.220
(These are the OpenDNS nameservers)


The output of zmlocalconfig | grep ldap:
(I noticed that this was asked for in another post). The reason that ldap is pointed towards localhost is because this box is firewalled, and that port is not available from an external interface.. Does ldap listen on an external interface?
ldap_cache_account_maxage = 15
ldap_cache_account_maxsize = 5000
ldap_cache_cos_maxage = 15
ldap_cache_cos_maxsize = 100
ldap_cache_domain_maxage = 15
ldap_cache_domain_maxsize = 100
ldap_cache_server_maxage = 15
ldap_cache_server_maxsize = 100
ldap_cache_timezone_maxsize = 100
ldap_cache_zimlet_maxage = 15
ldap_cache_zimlet_maxsize = 100
ldap_connect_pool_debug = false
ldap_connect_pool_initsize = 1
ldap_connect_pool_maxsize = 50
ldap_connect_pool_prefsize = 0
ldap_connect_pool_timeout = 120000
ldap_connect_timeout = 30000
ldap_host = steve.ac
ldap_is_master = true
ldap_log_level = 32768
ldap_master_url = ldap://localhost:389
ldap_port = 389
ldap_root_password = *
ldap_url = ldap://localhost:389
postfix_sender_canonical_maps = ldap:/opt/zimbra/conf/ldap-scm.cf
postfix_transport_maps = ldap:/opt/zimbra/conf/ldap-transport.cf
postfix_virtual_alias_domains = ldap:/opt/zimbra/conf/ldap-vad.cf
postfix_virtual_alias_maps = ldap:/opt/zimbra/conf/ldap-vam.cf
postfix_virtual_mailbox_domains = ldap:/opt/zimbra/conf/ldap-vmd.cf
postfix_virtual_mailbox_maps = ldap:/opt/zimbra/conf/ldap-vmm.cf
zimbra_class_provisioning = com.zimbra.cs.account.ldap.LdapProvisioning
zimbra_ldap_password = *
zimbra_ldap_user = zimbra
zimbra_ldap_userdn = uid=zimbra,cn=admins,cn=zimbra
zimbra_zmprov_default_to_ldap = false

I would also be interested to know if there is a way to get some more debug info in zimbra.log.

Your assistance in this matter would be greatly appreciated. Thanks.

[TMH]

An update! It seems that port 389 is not firewalled off as previously suspected.

[zimbra@vps ~]$ zmlocalconfig -e ldap_url=ldap://steve.ac:389
[zimbra@vps ~]$ zmlocalconfig -e ldap_master_url=ldap://steve.ac:389

However, this has not solved the problem.

[TMH]

Another interesting observation that I have made:

In my zmsetup.log, there are a number of messages that go along these lines:

Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.

I asked in ##java on freenode, and apparently this means the virtual machine has a set size of memory that there wasn't enough space for. This is odd, as I have 1gb of memory on this machine.

Could the above messages mean that the installation did not actually proceed properly? It would make sense, given that it is mentioning modules that I did not select.

[phoenix]

The usual reason for spell not working is a DNS problem, do you have correct DNS A & MX records set? Yuor /etc/hosts file is also not set-up correctly, review the Zimbra Quick Start Guide for details. You also mention nothing about the version of Zimbra, your operating system nor the hardware it's running on.

[TMH]

Oh, I am sorry, I should have thought of that. The server is a virtualised system- 1gb of RAM, 10gb HDD. The distro is CentOS4. I am attempting to install the latest open source edition of ZCS, the RHEL edition. CentOS is essentially a free version of RHEL as I understand it. The precise file that I downloaded was zcs-4.5.5_GA_838.RHEL4_64.tgz.

Here are the contents of my hosts file now. I think this is what it is meant to look like:

127.0.0.1 localhost localhost.localdomain
207.210.111.143 steve.ac


At the time, I was setting up ZCS without a working DNS server, and while I did this, spell was failing. I now have one, and it installs flawlessly.

I should actually have included my complete zonefile (I think that's what it's called):

$TTL 14400
@ 14440 IN SOA one.ns.steve.ac. steve.steve.ac. (
2006111705
14400
7200
1209600
86400
)

steve.ac. 14400 IN NS one.ns.steve.ac.
steve.ac. 14400 IN NS two.ns.steve.ac.

steve.ac. 14400 IN A 207.210.111.143
one.ns.steve.ac. 14400 IN A 207.210.111.142
two.ns.steve.ac. 14400 IN A 207.210.111.143

*.steve.ac. 14400 IN A 207.210.111.143

steve.ac. 14400 IN MX 0 steve.ac.

This precise record has worked fine the the last year or so with postfix, so I'm very sure that it works.

After reinstalling it again, the same issue still presents itself, and I noticed this error.

Initializing Documents...Could not create the Java virtual machine.


Thanks a lot for your help so far.

[TMH]

I am now attaching my zmsetup.log, maybe there will be some useful debugging info in here.

[phoenix]

You should be using a FQDN for your server, something like mail.steve.ac and the hosts file should contain

x.x.x.x mail.steve.ac mail

If this is behind a NAT device (?) you should point it to your LAN IP, check the wiki for details of the required set-up. If your not on split DNS those instructions should give you the necessaryu details of the required steps to fix it. You have a priority of 0 on your MX record, is that allowed?

[phoenix]

Originally Posted by TMH

I am now attaching my zmsetup.log, maybe there will be some useful debugging info in here.

The log indicates a DNS error resolving your MX (& probably A records), you'll need the records set-up to resolve this domain name.

[TMH]

Zimbra tries to autodetect my hostname, but it is incorrect, so I tell it that my hostname is steve.ac. I'm now using bosch.steve.ac. There are MX records and A records for my domain name.

The server is not behind NAT.

Here is my new hosts file:
127.0.0.1 localhost localhost.localdomain
207.210.111.143 bosch.steve.ac bosch

I've reinstalled again, and the problem still remains.

There are still plenty of messages that say "Could not create the Java virtual machine.". Are you sure this isn't at fault? Is there a way to get zimbra to give me even more debug info? Quite simply, there should be some log messages somewhere saying why other things are not starting up.

[phoenix]

What do the following show:

host `hostname`  <- backticks not single quotes
dig steve.ac 
dig steve.ac mx

Are there still the java VM errors in your zmsetup log? If there are you probably don't have enough RAM on your system (I know you said it's 1GB but that should be enough to get it running). Have you also disabled SElinux?

[edit] You can display the memory with the command 'free -m' - see what it says.