Problems with AFP (File Sharing) and Time Machine network Backups on OSX 10.7 Server

[jyhash]

Hi there folks,

I'm having an issue with the new 10.7 server. We're running 10.7.3 Server on an Late 2006 XServe with 10GB RAM , connected to a Drobo Pro for Time Machine Backups of our 50+ users via our internal network.

Recently, Time Machine backups have been getting delayed, stating that they can't connect up to the backup drive (which is out Drobo Pro handling 16TB of space). The Drobo Pro hass not unmounted, and all the IP Addresses coming from and going to the Server handling the backups will still ping.

We've also be having trouble with Server resident drives and shared folders that people can mount using the ⌘+K shortcut or "Connect to Server" option in finder. People will choose the IP address, and finder will think for a few minutes then give an error stating that it cannot connect the the specified IP. The only way I can reestablish connection to the file sharing through a remote computer is to remote or physically login to the server, then go into System Preferences and turn "File Sharing" off, then on again. After that, the ⌘+K on a remote machine can find the drives again.

It's driving me crazy why this is ocurring. I don't know why, and people in the office are beginning to complain that they can't access certain files on the servers or are worried because their backups are not completeing.

It should also be known that a majority of the machines we have in the ofice are still running 10.6.8, with only a handful on 10.7, as we haven't had the time to make the major switchover yet.

I'm not the best when it comes to server software or Unix CLI management (the office outsourced our Network Admin to the east coast), so I'm a bit stumped.

Does anyone have any suggestions to try? Is there a link between the number of Time Machine Users and the connection failures? Should I just tel the NetAdmin to create a script that will restart the File Sharing protocol on the server every couple of hours?

Hopefully someone will be able to help.

Thanks,

JYHASH

 

[kas23]

I am also having problems with Lion Server in terms of Time Machine. When you have Server loaded on your computer, no matter if everything is shut off, you will only be able to back-up to 1 single volume. No matter if you are trying to back 50+ people, they will only be able to back-up to one single volume or partition. Yes, that's friggin crazy. Due to this problem, I'll probably pull the plug on Lion Server.

 

[MacWorks]

AFP on 10.7.3 Server broken for me too

I'm having similar issues. I have a Drobo S connected to my server via FireWire and several of the folders on the Drobo S are shared via AFP. When connecting from another computer, I frequently only see the share points that live on the boot volume, but not on the Drobo. To get the Drobo's share points accessible again, I too have to turn File Sharing off and back on again.

 

[matspekkie]

I'm having similar issues. I have a Drobo S connected to my server via FireWire and several of the folders on the Drobo S are shared via AFP. When connecting from another computer, I frequently only see the share points that live on the boot volume, but not on the Drobo. To get the Drobo's share points accessible again, I too have to turn File Sharing off and back on again.

I was having similar issues as well. I got it fixed by removing groups and users permissions and then add them again. for some reason read write permissions got f... ed up. this was not on a time machine share though.

 

[skerfoot]

Abandon all hope

I fought with time machine and server for quite a while (there's a reasonably long thread about it here somewhere).

I eventually quite and decided to use Chronosync for backups instead. Connection is never an issue, you can backup however you want to and to what ever target(s) you want to.

It's not the apple way, and it's not (quite) as pretty, but it actually works and therefore better, in by books.

I used to run a system which had 10+ computers backing up to an OSX server using Chronosync, so I think it would scale up well for larger groups.