Wiki Service error message (continuous) in OS X Server (ML) 2.2

andrewnesbitt wrote:

Can you run each of these four commands individually and provide the output?

sudo serveradmin status wiki

sudo serveradmin status postgres_server

sudo ps aux | grep collab

sudo serverctl list

A.

Can I accomplish the results you seek from a REMOTE LOG IN, to the server?

H

andrewnesbitt wrote:

Can you run `sudo -s` to become root? Then you should be able to omit the sudo from each of the commands I gave you.

No. I get user already exists. Will restart help?

sudo: unable to cache user root, already exists

H

trilogy1000 wrote:

I tried running sudo psql -U collab -h "/Library/Server/PostgreSQL For Server Services/Socket" and got the following error:

psql: could not connect to server: No such file or directory

Is the server running locally and accepting

connections on Unix domain socket "/Library/Server/PostgreSQL For Server Services/Socket/.s.PGSQL.5432"?

I can connect to Postgres with PgAdmin but I'm assuming that's not the same as Postgres for Server Services. I tried to simply start the postgres_server but this also returned an error.

sudo serveradmin start postgres_server

postgres_server:error = "CANNOT_START_SERVICE_TIMEOUT_ERR"

I'm not sure where to go from here - any advice would be greatly appreciated.

The way I found with help from Apple Support, was to start over and import my services settings from the previous one by migrating them in. I used Migration Assistant to do it from a clean install of Server. It coincided to a new Mini upgrade at the same time, so it was a newer machine, upgrade to ML OS X 10.8.2, Server install and migration. That's really starting over. Since then, have had no errors. It was felt that a combination of corrupted files and a bad installation of files possibly with wrong permissions were not allowing the settings to be properly read. I proved that the Wiki errors did not exist on the new install before migration. They have not reappeared.

While I hesitate to recommend this to you, for my install this worked. You probably could try it. Good Luck!

Henry

joel2931 wrote:

Most likely the initialization of the postgres database failed for some reason when you installed Server. Otherwise its broken (or not running) for some other reason.

Could you post the contents of /Library/Logs/ServerSetup.log? That should answer several questions that I have. If postgres failed to initialize, a hint as to why should be found in that log. Was this a migration of data from another server or Time Machine backup? An in-place upgrade from an older server? A clean install on a system that hasn't had server installed on it in the past?

Also, had you installed a third party distribution of postgres on the system before installing server? We've found that this may cause problems.

Version 2.2.1 is available now and fixes several bugs related to postgres, if you have the option of doing a fresh installation.

Sounds like you are confirming what I just posted. Although I checked the version of my Server software and itthe original version 2.2.0 was fixed by starting over, it is clear to me that once broken postgres is pretty much not repairable. The update is on MAS at version you specified. Will update and try out. My takeaway on the issue, though has NOT changed. Once broken (Wiki or other errors happen) the reinstall of the entire OS should be done from a fresh copy and a migration must be done from a previous backup.

My question to you is, since I already did a fresh install and it's working on 2.2 now, do I need to do a full OS update as previously done for each update, to update the Server in place or will a simple update through MAS work? If not, why (since MAS makes no such distinction)?

I reinstalled and used the migration assistant. Everything migrated properly.

aptmxLS wrote:

Still have this issue and can not get it resolved no matter what I try.

Here's what I have done:

Trashed Server.app and entire Library/Server folder. Moved app back and re-built. Wiki Error reading settings message still there.

Did same as above but with trashing the Server.app and re-downloading a fresh copy. Same result.

Did the "last resort" of trashing the Server and Library/Server folder and ran the WipeDB script as per Apple's request. Had to set up all of Profile Manager again (no easy way to back up and restore Profile Manager database at all) and Wiki still states Error reading Settings!

Unfixable in my case?

You are far from the last resort my friend. Unfortunately, I don't know the complete history of what you did or didn't do in very much detail nor am I expert in Server. To me though, sounds like you are simply recreating your problems, whatever they are.

I happened on my solution by (happy) accident. Later (after my fix) installed the later version of Server (2.2.1). While it addresses those issues mentioned in this thread (e.g., keeps them from building up) and once later Server version is operational no longer builds up the problems, however, IT DOESN'T FIX THEM. Sounds like your case. The rub is in the fix. In my case it was easy to start over. Might not be for you. If not starting over, have no idea what your solution might be. Do know what I did and will share it with you again, from the beginning.

I had just gotten the Mini (a factory refurb Lion Server with included upgrade path to ML)...

1. As a hardware upgrade to the new equipment, I was able to speed it up with an SSD startup drive, add RAM and reuse the second HDD externally. So I had a pretty fast hardware with two internal drives but not ML and an old Lion Server which started up just fine at that point and gave no Server errors (still pristeen) at this point.
2. Next, did a fresh install of ML on the new SSD drive, keeping the old HDD drive intact.
3. Next, used Migration Assistant to bring in all data from a pre-existing Mini to the new start up drive hardware, which went exceptionally well in target mode. I was able to install a fresh version of ML, all programs, all settings using Migration Assistant (having failed with simply a backup restoration).
4. Then, installed (the then existing version of) Server (2.0?), then used Migration Assistant to bring in all settings (Server settings) from my old backup AS THEY WERE. Migration Assistant aparently was smart enough to know how to import the settings, where to import them AND change the permissions. It only imported the settings not the updates. I honestly think this is your key in Server.
5. This was explained to me by Apple Support and I did not realize until then how important this is in the case of ML Server. The other setting they had me change is to use "server.private" instead of "server.local" for my Server network setting. Bit of a chore logging in on the network, but I do it so rarely that it's no trouble.
6. Only after the old version of Server was working did I upgrade to the newer version(s). Thereafter not a lick of Wiki Server errors (I don't use Wiki Server at all) through two upgrades.
7. In Server the only services I use are Caching, DNS, File Sharing, Software Update.
8. My advice to you: Try it the Apple way.

HTH

aptmxLS wrote:

Turns out it was an issue with a program running on the Server called GoodSync. Worked with Apple Enterprise support and we finally figured out that a service was running called gs-server. This process was listening on port krb524, essentially blocking the process teamserver (Wiki) from listening. The gs-server process would overtake the teamserver process which caused the Error Reading Settings just to appear in the Wiki service. Uninstalled GoodSync and made sure the process gs-server was not active. All is working well now. I would have never figured that out. I knew it was something small like that. Hope this can help everyone else out who happens to use that program and runs the Server.app.

Just curious, how did you fnd this service running outside of and around Server? The console data is very difficult to understand.