Server Admin can't connect

Reading back through this entire thread, I seem to be having the issue spraguga was having, on an Intel Xserve running 10.6.2, but unfortunately, the "fix" of replacing the files from /System/Library/LaunchDaemons/ and then unloading/reloading the plist file hasn't worked, since I only get the first line in the servermgrd.log, not the second line that had been happening prior to the service stopping responding on 5/5/10.

Previously, the log read:

#Start-Date: 2010-04-25 18:53:09 PDT
#Fields: date time s-comment
2010-04-25 18:53:09 PDT Initalizing plugins
2010-04-25 18:53:16 PDT Listening for requests

Then, on the 5th, the log reads:

#Start-Date: 2010-05-05 18:05:30 PDT
#Fields: date time s-comment
2010-05-05 18:05:31 PDT Initalizing plugins

And with each "re-launch" of servermgrd after that, the second line, "Listening for requests" never appears, which seems to answer the question for me. In wondering, "Why isn't the service connecting?" it seems clear that this is not a DNS issue (all verified), not a firewall issue, not a switch issue (all unmanaged & verified, in my case), but a SERVICE issue.

Unfortunately, with the unload/load not working, I'm not sure how to get the service to listen again.

After a 24 hour 6.3 Snow Leopard Server installation session, I've found the following. Nos. 3, 5, 8, and 10 give Server Admin can't connect errors requiring a reinstall to fix. Nos. 1 and 2 don't even get far enough to open Server Admin. A "clean install" means installing to the second internal drive.

1. The in-place upgrade just fails. Not even worth attempting.

2. Doing a clean install and then using the migration assistant also fails.

3. Doing a clean upgrade and selecting any of the services during the install fails.

4. Doing a clean upgrade and selecting no services succeeds.

5. Any attempt to use open directory makes the rest of the system unusable.

6. Using NAT forces you to change your entire 192.168.X network. My 192.168.x network is not all that big but for many this would be a show-stopper.

7. Using NAT forces you to open every port on the NAT.

8. You can no longer run an RBL DNS zone. Doing so makes Server Admin lose it's connection and forces a reinstall to correct.

9. Apple Support wants $695 to support one problem.

10. The host name changes frequently on it's own. You can no longer set a hostname and a number of aliases and have the hostname stick. In some cases Snow Leopard Server chooses a hostname from the internet. There is no control over this, and if you attempt it a clean install is needed to recover.

11. Using the spam filter on the server means that 90% of your mail gets trashed by Dovecot.

12. Greylisting is provided silently. Unfortunately, this blocks a lot of legitimate mail. It's easy to turn off though and unlike DNS, editing the files the proper way (through the command line) doesn't trash the system.

That's all I've found so far.

I have this same problem on 6 xservers, including the brand new one I just configured.

After an initial connection locally on the new server, I can no longer connect with Server Admin tools.

I can however, connect via Server Preferences... (BUT ONLY LOCALLY)

Message was edited by: zmallory

Hi,

Like others here I am unable to run Server Admin, it loads but will not connect. I have tried all the suggestions in this thread and so far none have worked.

I am fortunate enough to have a Mac Mini server running as a back-up and alternate location for LDAP and Casper.

I have tried to connect to the primary server from the Mac Mini server, using the Server Admin tool but get the same result - no connection.

Hopefully Apple will come up with a solution ....soon.

Cheers
Phil

Had this problem happen to me yesterday.

lunchctl list | grep server did not include: com.apple.servermgrd

(and I didn't know about it by name, so I tried to muck with the next logical possibility: com.apple.RFBRegisterMDNS_RemoteManagement.server without success.

A reboot solved the issue and comp.apple.servermgr was loaded properly and server manager worked fine.

Same thing today on my Mac Pro running 10.6.3 Server, Casper 7.21, AFP, http, Tomcat. no servermgd processes at all, couldn't connect locally via IP, .local hostname, either FQDN, localhost, or 127.

Restart resolved the problem.

I'm puzzled now, during my attempts to resolve this I have restarted the server four times - so a simple restart did not resolve the issue for me.

I can't just keep bouncing the server without some explanation or guarantee that it will work.

Cheers
Phil

Happened to me again today. The first time (a couple of months ago) spraguga's fix worked for me. Today, nothing.

I have it happening to as of this morning. Macmini server wont load as well as on my Macbook pro 10.6.
Clicking on Server admin in the dock it takes an excruciatingly long time to load. Once it does I have the disconnected lightning bolt beside the server name.

The server runs fine and all the clients behave properly (I haven't heard anyone complaining yet).

I did add a MacMini 2.6 to the network last night before I headed home and we had a brutal power failure this morning that drained the batteries in the back up PS. That probably has nothing to do with it, but thought I would mention it if it made anyone's bells ring.

Add me to the list and DNS is configured right. All I know is the server has been on 20 days and nothing has changed.

I find it unbelievable that I have to restart the machine to (possibly) solve this problem.

Has anyone hear from apple about this issue?

I have 1 server running 10.6.3 which has the same problem.

Having similar issues here with a 10.6.3 server installation.

some points i tested and/or noticed:

a) Console.app -> servermgrd.log
right now contains

Initializing Plugins

but not the usually following line

Listening for requests

b) telnet hostname 311
does not work

c) http://server_ip:311
does not work

d) sudo ipfw list | grep 311
shows a correct ipfw rule

e) ps xau | grep servermgrd
shows a running process servermgrd -x (by root)

f) launchctl list | grep server
lists Screensharing.server & midiserver - but not servermgrd

g) netstat -na | grep LISTEN
shows 311 is not listening

h) console.app -> All messages
Servermgrd: error in getAndLockContext: flock(servermgr_info) FATAL time out
servermgrd: process will force quit to avoid deadlock
com.apple.launchd: com.apple.servermgrd Existed with Code 1

i) kill SERVERMGRD_PID
kills the daemon but autorestarts it by itself - still same output as mentioned in a)

j) and **** yeah ... its not a DNS issue here either - at least i am pretty pretty sure - while i cant prove it 😉

anyone reported that issue/bug in the meantime to Apple?


Best Regards
fidel