Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm testing an Amanda install on an ubuntu 10.04 server as a backup server for primarily OS X 10.6 clients.

The server has been setup and has successfully run backups on specific local directories for about a week. I've installed and configured the OS X client according to the Zmanda wiki article. I don't have an internal DNS server in this test environment, so in the disklist entry on the server I defined the client by IP and not FQDN (ie. 10.10.10.5 /Users/admin test-home-tar).

When I run amcheck on the server, this is the output:

WARNING: 10.10.10.5: selfcheck request failed: Connection timed out
Client check: 2 hosts checked in 73.001 seconds. 1 problem found.

What am I missing to get the client and server to successfully communicate and run a backup?
Do client disklist entries need to be by FQDN?

share|improve this question

There are a number of things that can cause this error. To answer the simple question, no, the DLE entry does not need to specify clients by FQDN.

Amanda uses an rhosts-like configuration on the client to control what hosts are allowed to speak with the client's amandad service to initiate transfers. If the client's .amandahosts file lists the Amanda server by the server's FQDN, then the client must be able to reverse resolve the IP of the incoming connection and match the resolved name against the FQDN in the.amandahosts file. If it can't do this, the client will reject the incoming connection. You can work around this with crafty entries in the client's /etc/hosts file, but it may be simpler to just list the server's IP address (as viewed by the client) in the client's .amandahosts file instead. Even this can get tricky if the client and server have a firewall between them, as the client will probably see the IP of the firewall on the incoming connection, not the server's own IP.

Keep in mind that this is only one of a number of issues that can cause this error. A list of potential causes includes:

  • server can't ping the client or some other basic networking issue between them.
  • Client's inetd not configured to launch amandad for incoming 10080 connections
  • Client's inetd not listening on the correct interface
  • Client's kernel firewall not allowing port 10080 traffic from the server (iptables, ufw, etc.)
  • Server's firewall is not allowing returning packets from the client
  • Client's .amandahosts file missing, or ownership or permissions incorrect
  • tcpwrappers issue (/etc/hosts.allow, /etc/hosts.deny) on the client
  • External firewalls between client and server
  • DNS resolution conflicts if client and server use different DNS resolvers
  • Client's amanda-client.conf file specifies a different authentication method than the server's DLE requires (BSDTCP vs SSH, for example.)
  • Client's Amanda software is elderly and doesn't even support BSDTCP or SSH. (So use BSD instead.)
  • Server has multiple nics or IP addresses and it's using a different IP than the client expects.

and so on.

See this link to the Amanda Wiki for more troubleshooting tips for this common error message: https://wiki.zmanda.com/index.php/Selfcheck_request_failed

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.