Server Admin can't connect

Hello all !

Well, I have this problem too: server admin can't connect to the server.
Tried to connect to server.local, 127.0.0.1, his IP address, and nothing.

Also, on terminal, telnet 127.0.0.1 311 I have a connection refused.
lsof -i :311 shows me nothing (Of course, because no more listening on port 311)

Spraguga's hint did not work either. But I still wonder how he can use the command line sudo launchctl unload -w /System/Library/LaunchDaemons/com.apple.servermgrd.plist if he deleted this file. If I do that, I have an error message. So I just switched his 2 first steps. I used the .plist from another server, but without good result.

I prefere not to restart the server. All services (mail, web, ftp and afp) work as usual (no problem).
This server is connected to an OD Xserve.

Just for information: on this server, I worked about the vacation notices that does not work. Seems to be related to some kind of crash of the com.apple.emailrules.plist. So I tried to do a launchctl unload / load of this plist. It gave me confirmation of this problem, but (maybe because I forgot to close ServerAdmin, Server Preferences and Workgroup Manager), now I'm unable to launch them except WGM.

Any way to activate the ports related to these services ?

I've been having the same problem this morning on a mini server (brand new) running 10.63. It started after making some firewall settings. I already had the firewall active and working fine, then added the multicast port. Then the admin started spinning endlessly. I rebooted and there was no connection. Rebooted again and waited. The connection eventually came back after about 10 minutes, but no settings available in Server Admin. 10 more minutes of waiting and everything came back. For the heck of it (I am a masochist I guess) I attempted my changes on the Firewall again. This time I let it spin for the 10 minutes and it was ok. Very strange.

Sorry, I didn't correctly read this part... I searched alot and got my eyes really tired 🙂

Just that I can't find any com.apple.servermgrd.plist in /System/Library.
I suppose it's /Library/Preferences/

Will check that and will come back later.

<knock hardly his head on the wall>

Still no luck... Will try to restart the server...

Anyway:

ps auxww | grep servermgrd
root 50000 0.0 0.4 2494700 13400 ?? Ss 5:20PM 0:00.10 servermgrd -x

So, it's running

netstat -na | grep LISTEN
=> there is no port 311 on which the server is listening...

In the error logs, I just have this each time I want to open Server Admin:

2010-04-16 17:20:04.776 servermgrd[50000:207] -[AccountsRequestHandler(AccountsOpenDirectoryHelpers) openLocalLDAPNodeIfNeeded]: dsLocalLDAP = (null), error = Error Domain=com.apple.OpenDirectory Code=2000 UserInfo=0x1019009e0 "Unable to open Directory node with name /LDAPv3/127.0.0.1."

Firewall is not activated, even if ipfw -i gives me this:
12306 allow tcp from any to any dst-port 311

Other rules don't interfere with this one.

Did not work for me. I don't know if it was a 10.6.2 thing but I also had to chown the file to get it to load up after copying it over from a good machine.

I Unloaded the servermgrd, replaced the file with a known good servermgrd (chowned it to root), loaded it. In the servermgrd.log I get:

#Start-Date: 2010-04-16 11:12:02 PDT
#Fields: date time s-comment
2010-04-16 11:12:02 PDT Initalizing plugins
2010-04-16 11:12:05 PDT Listening for requests

But when I try the app I get the Cannot connect still and doing a port scan on 127.0.0.1 from the machine gives me:

Port Scanning host: 127.0.0.1

Open TCP Port: 88 kerberos
Open TCP Port: 139 netbios-ssn
Open TCP Port: 445 microsoft-ds

So something is still messing up port 311. Gonna rebuild this server this weekend I guess, but this is a really annoying bug.

Still didn't work, kicked it up to 10.6.3, repaired perms, tried it again but still no port 311.

I rebuilt the server with the same IP, same location on the switch, same everything and now port 311 is listening correctly. Pretty drastic bug fix, fortunately all this guy did was serve files, guess I got off easy.

The 10 minute wait rings a bell for me. I have no idea if this is what some of you are running up against, but this is a TCP stack forced wait time (TIMED_WAIT) and it is causing a local server I wrote for my local household LAN some grief. Turns out for TCP to work a server port isn't considered free for reuse for 10 minutes after it is closed. (For the gory details why this is so see section 2.7 at <http://unlser1.unl.csi.cuny.edu/faqs/sock-faq/html/unix-socket-faq-2.html>.) This is enforced by the kernel TCP stack and I've not come across any way to inhibit or reduce this timeout behavior. If you took more than 10 minutes doing something to "fix" the problem perhaps you didn't fix it but unknowingly permitted this timeout to occur.

To see if your port of interest is in this forced wait (10 minutes on my Snow Leopard) do 'netstat' in terminal or via the Network Utility. If your port shows the dreaded TIME_WAIT then - bingo. Here's what mine looks like:

Active Internet connections
Proto Recv-Q Send-Q Local Address Foreign Address (state)
tcp4 0 0 10.0.1.2.62351 10.0.1.2.62510 TIME_WAIT

Once again I'm data mining the net to see if there is any way to fix this. All packets transfer in milliseconds on my local LAN and I have no LAN routers or bridges so having to wait 10 minutes every time I restart my server is absurd. If any of you find a solution please post about it here. I'll do the same.

Good Luck

hi there

I've been having the same issues, and to my surprise, after the 3 reinstall (and me basically giving up) I left the machine over night and then reconnected using serveradmin in the morning and it worked. I really have no idea what was causing it. I tried all suggestions on this thread as I had also blitzed the plists in desperation. What I can say is that this problem definitely occurs when installing the 10.6.2 update. worked fine from 10.6.0 to 10.6.1. I am yet to go to 10.6.3 but suspect I will hit the same thing again. Incidently when I have tried the 10.6.3 combo update it fails to install.

I am convinced this is DNS behaviours, and sorry to ask this question in this thread, but think it could help some people out there that hit this problem on osx server and the macmini (as they are probably like me a home business that has the machine running inside a network as the only server publishing a few services to the outside world): 'Why do I need DNS running as a service just to get a fqdn for the box? Can't I just drop an fqdn file in somewhere?' I seem to remember doing this on unbuntu at somepoint in the past...

Anyway - in summary, waiting helped - it sorted itself out without a reboot..

Hi

+"I am convinced this is DNS behaviours"+

Poorly configured/implemented DNS would cause general server instability anyway. Not a surprise then if there are problems trying to use the Server Administration applications locally or remotely.

There are many that don't fully understand the reliance OSX Server and its technologies place on DNS. However this was not the OP's reason for posting. His problem was due to Server Admin not listening anymore on port 311. His DNS was configured as it should be. Having contributed to this thread initially and having said I'd never seen the problem, sod's law intervened and I saw exactly the same problem the other day.

This too was a fully patched Server. Absolutely nothing unusual in the logs. The cure? Kill the servermgrd process, unload the com.apple.servermgrd.plist, restart the server, reload the launchd item. Not a problem thereafter. This was not the same 'cure' as OPs. Clearly there are some differences somewhere which makes this problem all the more frustrating to troubleshoot and fix.

HTH?

Tony

Today I got the same problem. Everything on the server (10.6.3) worked, but I was unable to administer the server via server admin tools - no matter if I tried via network or local.

In my case the solution seemed to be a missconfigured apache. I had added a new site in the etc/apache2/sites/ folder through ssh, therefore I didn't notice that the admin tools did not work any longer.

I just removed the XYZsite.conf and now everything works fine.

Hope this helps - good luck!
Greetings
Sven

Message was edited by: Sven Koesling (typos... now hopefully gone 🙂 )

hm, sry, I posted to early.

I didn't do anything on the server, now server admin tools cannot connect anymore - not local, not via network.