Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

Fresh Leopard install grinds to a halt - please help!

Hi all,

I'm having some serious problems with Leopard server on a new Xserve and this latest issue I can't figure out. Fresh install of Leopard, connected to Active Directory, running the AD-OD 'magic triangle'. All that is working well.

However, the longer the server is up for, the slower it becomes - and this is without anyone using it! Over the course of a day, shell access in the terminal from my admin machine goes from being instantaneous to sluggish, then after returning from the weekend there was no response after 20 minutes of trying. DHCP service also seemed to be dead - no new addresses were being handed out. Server had to be rebooted and then access was instant again.

I get the feeling that there's something not quite right with my network setup, but I don't know what. The server is connected by link aggregate (4x 1Gbps connections) to the local network, while the two ports on the Xserve motherboard are connected to the university system for internet access. I used the Gateway Setup Assistant to configure the networking (internal 192.168.1.x) for NAT, DHCP, Firewall, VPN, DNS...

Looking back through the log files, this is what starts to appear:

Dec 5 14:43:21 mh-novst01 com.apple.DirectoryServices[25]: Enter machine password:
Dec 5 14:43:21 mh-novst01 com.apple.KerberosAutoConfig[23069]: Kerberos configuration is up to date
Dec 5 14:43:37 mh-novst01 bootpd[23078]: re-reading /etc/bootptab
Dec 5 14:43:37 mh-novst01 bootpd[23078]: Loaded 0 entries from bootptab (0 bad)
Dec 5 14:43:38 mh-novst01 com.apple.DirectoryServices[25]: DNS update failed!
Dec 5 14:46:40 mh-novst01 natd[142]: failed to write packet back (No route to host)
Dec 5 14:46:42 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:46:49: --- last message repeated 1 time ---
Dec 5 14:46:49 mh-novst01 sshd[21328]: DEAD_PROCESS: 21342 ttys001
Dec 5 14:46:53 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:47:10: --- last message repeated 7 times ---
Dec 5 14:47:10 mh-novst01 natd[142]: failed to write packet back (No route to host)
Dec 5 14:47:28 mh-novst01 /usr/sbin/ocspd[23139]: starting
Dec 5 14:47:50 mh-novst01 /Applications/Server/Server Monitor.app/Contents/MacOS/Server Monitor[23084]: [23084] ServerManager session failed in connect(130.88.61.189,130.88.61.189,311): 60
Dec 5 14:47:54 mh-novst01 natd[142]: failed to write packet back (No route to host)
Dec 5 14:47:55 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:48:19: --- last message repeated 21 times ---
Dec 5 14:48:19 mh-novst01 natd[142]: failed to write packet back (No route to host)
Dec 5 14:48:20 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:48:44: --- last message repeated 19 times ---
Dec 5 14:48:44 mh-novst01 natd[142]: failed to write packet back (No route to host)
Dec 5 14:48:45 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:49:02: --- last message repeated 17 times ---
Dec 5 14:49:02 mh-novst01 Server Admin[20075]: * -[GroupTextField windowDidResignKey:]: unrecognized selector sent to instance 0x4d0be60
Dec 5 14:49:02: --- last message repeated 1 time ---
Dec 5 14:49:02 mh-novst01 ARDAgent [1495]: Front Process:Couldn't get front process.. error: -600.
Dec 5 14:49:02 mh-novst01 natd[142]: failed to write packet back (Host is down)
Dec 5 14:49:03 mh-novst01 loginwindow[550]: DEAD_PROCESS: 0 console
Dec 5 14:49:03 mh-novst01 shutdown[23169]: reboot by andy:
Dec 5 14:49:03 mh-novst01 com.apple.loginwindow[550]: Shutdown NOW!
Dec 5 14:49:03 mh-novst01 shutdown[23169]: SHUTDOWN_TIME: 1196866143 135837
Dec 5 14:49:03 mh-novst01 com.apple.loginwindow[550]: System shutdown time has arrived^G^G
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: Stopping IP Failover services
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: Disabling Network Address Translation
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: cat: /var/run/failoverd.pid: No such file or directory
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: cat: /var/run/heartbeatd.pid: No such file or directory
Dec 5 14:49:03 mh-novst01 SystemStarter[46]: IP Failover (23172) did not complete successfully
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
Dec 5 14:49:03 mh-novst01 com.apple.SystemStarter[46]: net.inet.ip.forwarding: 1 -> 0
Dec 5 14:49:03 mh-novst01 SystemStarter[46]: The following StartupItems failed to properly start:
Dec 5 14:49:03 mh-novst01 SystemStarter[46]: /System/Library/StartupItems/IPFailover
Dec 5 14:49:03 mh-novst01 SystemStarter[46]: - execution of Startup script failed


Looking back at the weekend logs I'm confronted with a whole load of:

Dec 10 13:28:56 mh-novst01 mDNSResponder[30]: Could not write data to client 24 after 60 seconds - aborting connection
Dec 10 13:28:56 mh-novst01 Unknown[30]: 24: DNSServiceCreateConnection
Dec 10 13:30:09 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:30:17 mh-novst01 natd[154]: failed to write packet back (Host is down)
Dec 10 13:31:58 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:33:34 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:35:31 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:38:01 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:40:51 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:43:16 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:43:16 mh-novst01 natd[154]: failed to write packet back (Host is down)
Dec 10 13:43:46: --- last message repeated 1 time ---
Dec 10 13:45:31 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:45:46 mh-novst01 natd[154]: failed to write packet back (Host is down)
Dec 10 13:46:16: --- last message repeated 1 time ---
Dec 10 13:48:16 mh-novst01 natd[154]: failed to write packet back (No route to host)
Dec 10 13:48:16 mh-novst01 natd[154]: failed to write packet back (Host is down)
Dec 11 00:00:00 mh-novst01 newsyslog[12484]: logfile turned over


Does anyone have any clues? There's a DNS system running on the university network, which is what the DHCP clients are pointing towards, but of course this will have no knowledge of our internal 192.168.1.x network. Should I be configuring the server in the DNS service in Server Admin? Or is this looking like an entirely different kind of problem?

Any suggestions would be very gratefully received!

Thanks,

Andy.

iMac 2.4GHz 24", Mac OS X (10.5.1)

Posted on Dec 11, 2007 1:25 AM

Reply
2 replies

Dec 11, 2007 3:55 AM in response to andydvsn

Update...

Thankfully I'm in a position to experiment with this at the moment, and on reinstalling OS X 10.5 Server I've discovered that the problems seem to start as soon as I turn on the internal network interfaces. Even without the link aggregate, just configuring and enabling one interface with an IP other than the one handed out by the university DHCP service causes Server Admin to freeze up and never successfully display the server details. Making the interface inactive brings it back to life.

The internal settings are very simple - just the usual; IP 192.168.1.x, Netmask 255.255.255.0... Am I missing something obvious here? I'm thinking there might be a DNS issue, or the service order in the Network preferences?

TIA,

Andy.

Jan 28, 2008 8:20 AM in response to andydvsn

Update #2...

I've managed to come up with a workaround for all these problems, although it's a bit long-winded. I shifted our network onto the 172.16.10.* range (didn't help the problem, just better practice as I'm running a VPN too) and pointed the client machines to look at our gateway box for DNS resolution.

I then configured BIND on the gateway to use OpenDNS servers, but revert to using the university DNS service should any requests come in for the Active Directory domain (using the primary / secondary DNS features). This prevents the lock-ups, but retains the ability to connect to AD correctly when required.

All working very well so far - a whole classes of AD / OD mixed users are logging in simultaneously and not having any problems.

Hope this helps someone else out! Please ask if you require any further details.

Andy.

Fresh Leopard install grinds to a halt - please help!

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple ID.