One of our servers had the DHCP service removed a while ago. However the system log keeps reporting "natd[199]: failed to write packet back (No route to host)", I'd assume because DHCP isn't there anymore. Trying to kill natd took the server offline completely.
Does anybody know why natd is still running and whether it needs to? What is the best way to remove/disable it?
I don't think the natd and DHCP issues are related. That natd message occurs when natd receives a packet that it cannot forward for various reasons. It doesn't depend on whether DHCP is active or not (either as a client or a server).
It could mean a misconfiguration in your NAT or firewall setup. The real issue is how frequently is it happening (e.g. once a day? twice a second?) and whether or not you expect/want to run NAT on this machine (if not just turn off the NAT service in Server Admin).
DHCP server and NAT had been turned off and removed from the server via Server Admin as not necessary anymore. Nevertheless the error message kept being logged about 3 times/min.
In the end I cured it by reinstalling DHPC+NAT, and then disabling (but not removing) them. This stopped the errors in the log files, but I'd have preferred a more elegant solution.
I've got the same or a similar problem getting "failed to write packet back (No route to host)" over and over.
How did you reinstall DHCP and NAT? Did you do a reinstall of Snow Leopard Server? Or did you turn off these services with Server Admin, then turn them back on?
I have a 10.5.8 Server that has the NAT/DHCP/Firewall hiccup-misconfiguration of the nature....
Here's the log
4/27/10 7:50:28 AM natd[317] failed to write packet back (No route to host)
4/27/10 7:50:28 AM natd[317] failed to write packet back (No route to host)
4/27/10 7:50:30 AM natd[317] failed to write packet back (Host is down)
4/27/10 7:50:30 AM natd[317] failed to write packet back (Host is down)
4/27/10 7:50:34 AM natd[317] failed to write packet back (Host is down)
4/27/10 7:50:34 AM natd[317] failed to write packet back (Host is down)
4/27/10 7:50:42 AM natd[317] failed to write packet back (Host is down)
4/27/10 7:50:42 AM natd[317] failed to write packet back (Host is down)
4/27/10 8:01:22 AM bootpd[4841] can't open /etc/bootptab
4/27/10 8:01:22 AM bootpd[4841] server name myLeopardServer.myDomain.com
4/27/10 8:01:22 AM bootpd[4841] interface en0: ip 10.0.3.100 mask 255.255.255.0
4/27/10 8:01:22 AM bootpd[4841] interface en1: ip 10.0.4.1 mask 255.255.255.0
4/27/10 8:01:22 AM bootpd[4841] DHCP REQUEST [en0]: 1,0:1e:c2:d1:e2:c8
4/27/10 8:01:22 AM bootpd[4841] ACK sent Anybody's iPhone 10.0.3.40 pktsize 318
4/27/10 8:01:22 AM bootpd[4841] ACK sent Anybody's iPhone 10.0.3.40 pktsize 318
4/27/10 8:11:54 AM bootpd[4873] can't open /etc/bootptab
4/27/10 8:11:54 AM bootpd[4873] server name myLeopardServer.myDomain.com
4/27/10 8:11:54 AM bootpd[4873] interface en0: ip 10.0.3.100 mask 255.255.255.0
4/27/10 8:11:54 AM bootpd[4873] interface en1: ip 10.0.4.1 mask 255.255.255.0
4/27/10 8:11:54 AM bootpd[4873] DHCP REQUEST [en0]: 1,0:22:41:75:6a:69 <iPod-touch>
4/27/10 8:11:54 AM bootpd[4873] ACK sent MyCompany's iPod Touch 10.0.3.4 pktsize 318
4/27/10 8:11:54 AM bootpd[4873] DHCP DISCOVER [en0]: 1,0:22:41:75:6a:69 <iPod-touch>
4/27/10 8:11:54 AM bootpd[4873] OFFER sent MyCompany's iPod Touch 10.0.3.4 pktsize 318
4/27/10 8:11:55 AM bootpd[4873] DHCP REQUEST [en0]: 1,0:22:41:75:6a:69 <iPod-touch>
4/27/10 8:11:55 AM bootpd[4873] ACK sent MyCompany's iPod Touch 10.0.3.4 pktsize 318
4/27/10 8:11:55 AM bootpd[4873] ACK sent MyCompany's iPod Touch 10.0.3.4 pktsize 318
4/27/10 8:11:55 AM bootpd[4873] ACK sent MyCompany's iPod Touch 10.0.3.4 pktsize 318
4/27/10 8:13:55 AM mDNSResponder[32] DNS Message too short
4/27/10 8:13:55 AM mDNSResponder[32] DNS Message too short
4/27/10 8:14:23 AM bootpd[4873] DHCP REQUEST [en0]: 1,0:1e:c2:d1:e2:c8
4/27/10 8:14:23 AM bootpd[4873] ACK sent Anybody's iPhone 10.0.3.40 pktsize 318
4/27/10 8:14:23 AM bootpd[4873] ACK sent Anybody's iPhone 10.0.3.40 pktsize 318
4/27/10 8:20:49 AM natd[317] failed to write packet back (No route to host)
4/27/10 8:20:49 AM natd[317] failed to write packet back (No route to host)
4/27/10 8:20:51 AM natd[317] failed to write packet back (Host is down)
4/27/10 8:20:51 AM natd[317] failed to write packet back (Host is down)
This pattern repeats and repeats on a newly rebuild 10.5.8 Server. Could this be caused by a Airport Base Station Gateway/DHCP and Leopard Server DHCP fighting on the same subnet? I know this is a silly question but I've had my AirportBaseStation doing port forwarding and serving as the gateway router with a LeopardServer as the target for most of the forwarded traffic.... This server has two interfaces serving as another NAT Gateway for my private IP space on a different subnet.
Can anyone tell me it that could cause this problem and if so, how can I prove to myself this is the issue? I'd like "proof" because this has been my general setup for this double NATed network for a few years and I think it works.... It has worked pretty well if not perfectly.... I have a fair but not absolute confidence that this works but I can't say that the log entries are not because of DHCP fighting....
I believe there is a problem but the configuration of something.....