Re: impending update of amanda packages to 2.5.2p1

[Urban Boquist <urban%boquist.net@localhost>]

>>>>> Greg Troxel writes:

Greg> I have done the update.  Please let me and Nick know of any
Greg> problems.  The update does not intend to do anything other than
Greg> move to 2.5 and associated necessary changes - other aspects of
Greg> the package should be preserved.

Hi Greg,

Thanks a lot for doing this upgrade!

Unfortunately I can't get it to work... ;-)

I've had a working setup for a long time with a server running amanda
2.5.2p1 (built by hand) serving a number of clients running 2.4.4p4
(built from NetBSD pkgsrc). No problems at all.

Now I upgraded one of the clients to 2.5.2p1 built from pkgsrc, and
the backup for that client now fails (the other clients still work).

FAILURE AND STRANGE DUMP SUMMARY:
  current  root  lev 1  FAILED [cannot read header: got 0 instead of 32768]
  current  root  lev 1  FAILED [too many dumper retry: "[request failed: 
timeout waiting for ACK]"]
  current  root  lev 1  FAILED [cannot read header: got 0 instead of 32768]

I can see on the client that it runs gtar and sendbackup but somehow
it seems to fail communicating the results back to the server. I don't
understand most of what is in all the /tmp/amanda files but in one of
the dumper files on the server side I see:

dumper: time 3.429: connect_port: Try  port 0: Available   - 
dumper: time 3.481: connect_portrange: connect from 0.0.0.0.64868 failed: 
Connection refused
dumper: time 3.481: connect_portrange: connect to 192.168.1.29.65498 failed: 
Connection refused
dumper: time 3.481: stream_client: Could not bind to port in range 0-0.
dumper: time 3.482: security_seterror(handle=0x805c400, driver=0xbbb783a0 (BSD) 
error=can't connect stream to current.boquist.net port 65498: Connection 
refused)

and a little later:

dumper: time 48.541: security_seterror(handle=0x805c200, driver=0xbbb783a0 
(BSD) error=timeout waiting for ACK)
dumper: time 48.541: security_close(handle=0x805c200, driver=0xbbb783a0 (BSD))

on the client side in one amandad file:

amandad: time 33.579: stream_accept: timeout after 30 seconds
amandad: time 33.580: security_stream_seterr(0xbb942000, can't accept new 
stream connection: No such file or directory)
amandad: time 33.580: stream 0 accept failed: unknown protocol error
amandad: time 33.580: security_stream_close(0xbb942000)
amandad: time 63.588: stream_accept: timeout after 30 seconds
amandad: time 63.589: security_stream_seterr(0xbb94b000, can't accept new 
stream connection: No such file or directory)
amandad: time 63.589: stream 1 accept failed: unknown protocol error
amandad: time 63.589: security_stream_close(0xbb94b000)
amandad: time 93.597: stream_accept: timeout after 30 seconds
amandad: time 93.598: security_stream_seterr(0xbb954000, can't accept new 
stream connection: No such file or directory)
amandad: time 93.598: stream 2 accept failed: unknown protocol error


The machines are on the same subnet, there are no firewalls or other
things like tcpwrappers that I can think of. And the same client works
fine if I go back to 2.4 again.

I took a tcpdump too, there is some UDP talk first but then the first
TCP setup seems to fail:

13:31:54.838900 IP amanda.boquist.net.64842 > current.boquist.net.65472: S 
3385844345:3385844345(0) win 32768 <mss 1460,nop,wscale 
0,sackOK,nop,nop,nop,nop,timestamp 0 0>
13:31:54.867041 IP current.boquist.net.65472 > amanda.boquist.net.64842: R 
0:0(0) ack 3385844346 win 0

Does anyone recognize this or have any good suggestions how to proceed?

Thanks for any hint,

        -- Urban