Q: For all of you having services failing to start

I had installed 10.7 Server with the hope of setting up a home server.  For many reasons, I have downgraded to SL and dropped the idea of my own server.

Just thought you all would want to know,  I think unless you are doing this for business, just stay away from Lion and its Server altogether and life will be much easier again.

I plan on testing out a Linux package to run inside a VM for my home server use and will probably never go back to a Mac server.  What's the point unless you run a network with only Apple devices?

Something new...

I noticed in the logs devicemgr having problems writing it's pid file in /var/run/devicemgr. Thing is, the folder didn't exist. I created it with the right permissions.

drwxr-xr-x  12 _devicemgr  _devicemgr   408B Feb 21 17:30 devicemgr

Than, I switched "reserve ressources" off, rebooted, stopped all services, did a "serveradmin stop" on all services to be shure (including serveradmin stop postgres), rebooted and restarted all services.

Tomorrow night/this weekend I'll be fixed on if it solved my problem.

Sadly amazing - this is a 'known' issue in the community for 8 months, and Apple hasn't found a fix yet. I'm disheartend. Working on a fresh install myself.  I even went as far as a TimeMachine restore, from pre-install.

I was testing on my own machine before rolling out to a few clients and upselling for Apple over Linux.  Not about to put my reputation on the line for this version of Server.  Sorry guys - it's about as reliable as a drunk clown.

C'mon Tim - Steve would have jumped in a long time ago to get this fixed. You can do better.

I think I've got a solution! I've got a week of uptime and everything's running good.

As root :

cp /var/db/dyld/dyld_shared_cache_i386 /var/db/dyld/dyld_shared_cache_x86_64 /var/run/

update_dyld_shared_cache -force

I just don't know if the problem comes back after a reboot, since I didn't reboot after that

Didn't work for me, unfortunately.

Problem came back after a reboot. Someone suggested a safe boot to solve the problem

http://support.apple.com/kb/HT1564

I've got things working for the time being with safe reboot, a permissions fix, & a reboot back into normal usage. Unfortunately, I can't run  in safe mode for normal use because I run  additional services.

It took a few goes before everything was happy. The first time made things worse; Server and Server Admin reported nothing, although ps showed services were running. That turned out to be https://discussions.apple.com/thread/3193142?start=15&tstart=0 --- the suggested fix worked. A further attempt got things back to where started. But the final go got everything working. I think you'll all understand why I'm not rebooting again, at least not until I absolutely have to.

I've had two more postgres failures along the lines of this topic (on different machines), and in each case a permissions check or two in safe mode, then a return to normal usage, resolved the problem, if temporarily. I don't know if this will keep things working, obviously, but it seems to help.

I just hope Apple fix the root cause.

I haven't been able to get serveradmin to manage postgres for a while now.  What I've ended up doing is manually start the postgresql service after a reboot and that seems to allow all the services to work properly.

The command line I'm running under sudo from a shell script is:

-------------------- Start of shell script ------------------

#!/bin/bash

nohup sudo -u _postgres /usr/bin/postgres -D /var/pgsql -c "listen_addresses=*" -c "log_connections=on"  -c "log_directory=/Library/Logs" -c "log_filename=PostgreSQL.log" -c "log_line_prefix=%t " -c "log_lock_waits=on" -c "log_statement=ddl" -c "logging_collector=on" -c "unix_socket_directory=/var/pgsql_socket" -c "unix_socket_group=_postgres" -c "unix_socket_permissions=0770" &

--------------------- End of shell script -------------------

The nohup line is all one line, no breaks.

Hope this helps anyone who just can't get it to start no matter what.

serveradmin stop postgres tries to stop in this case but fails and leaves things in a funky state, so you will have to manually stop/kill postgres if you need to restart it due to changing permission files, etc. internal to postgreSQL.

Thanks a lot David this is the one solution for me and works absolutly perfect!

Thank you. This solves my PostgreSQL- Problem on my Server (10.7.3).

If have opened an Ticket to Apple vor this Problem.

Hi guys,

I've been having this problem since I "upgraded" to Lion Server. The problem disappeared when I removed the DNSUpdate utility (dyndns lookalike), and appeared again when I reinstalled the utility to keep my dns ip address in sync.

Now, this might just be a coincidence. Does anyone having this issue also have DNSUpdate installed ?

Thanks for your input and good luck with this chaotic piece of software.

Arach

I have 2 Lion servers, only 1 has this service failure issue and both run almost the same services.

The original post from TheTeknologist (July 2011) solved my problem immediately.

Thanks a million!!!!

I can't believe that Apple have not fixed this yet!!

Ok, I got my problem solved, for now.

It all started with an "innocent" Software Update in Lion Server. Everytime I do a software update with this thing I just pray nothing breaks. Sure enough, on reboot, no more screensharing option listed in mounted server in Finder. Just Connected As option.

So I fired up Screen Sharing App... got in from there thankfully. However, Screen sharing has broken on me several times. Grrrrrh.

So then I noticed that my Wiki was causing this Curl Index error as mentioned in this thread. I did the following:

1. Repaired permissions, rebooted. I actually rebooted a few times before this to no avail. After about 20 minutes (subsequent to the permissions repair and reboot), the screen share icon returned to the mounted Mac Mini on my Computer. By the way, a ton of permissions were messed up.

2. Tried a bunch of stuff mentioned in this thread to no avail.

3. Turned Wiki Services on and off in the Server App. Voila! My Server splash page came back.

About me: Mac Mini latest, Lion Server 10.7.3, Dyndns pointing to a dynamic IP.

After seven months, I finally received an answer from Apple Enterprise Escalation Support Engineering. I don't know if this applies to everyone, but there were several undocumented fixes in 10.7.3 that were not mentioned in the release notes regarding Time Machine and PostgeSQL that ultimately resolved the issue for me.

According to Apple: 10.7.3 specifically resolves an issue with Time Machine backup, where PostgreSQL databases could become unusable because of an issue with backup and automatic recovery of Postgre data

Unfortunately to see the benefit I had to reinstall the system software. I had reinstalled the system software so many times before, that I resorted to using a boot disk created according to the methods described by MacWorld, et. al. to save time. If you run and upload an EDC to Apple they can tell how the system software was installed. As such they consider this an unsupported method and really are resistant to helping you until you re-install via a supported method. Having to reinstall the system software for a server implementation is really the nuclear option but I was willing to give it one more try. To make sure everything was supported, I had the techs at the Genius Bar install 10.7.3 from an image at the Apple Store. It was great in that it only took about 5 minutes to install versus the several hour process at home given the speed of my Internet connection. I'm happy to report that for the first time since October 2011 Lion Server is running flawlessly. It's been a week and so far and all my services come up as expected even after 5-6 reboots.