On 3/23/2013 4:01 PM, Glenn English wrote: > I installed wheezy on one of the LAN hosts yesterday. Everything else is > squeeze, including my amanda backup machine. Amanda kept saying my wheezy box > wasn't working: > >> WARNING: sbox.slsware.lan: selfcheck request failed: timeout waiting for ACK > > I looked (in man and google) for an explanation of this, and found references > to amanda-client.conf. There is no amanda-client.conf anywhere on my system, > even on the squeeze boxes where amanda has been working fine for years (since > sarge, IIRC).
This suggests the defaults worked for you and you simply never needed to create/modify the conf file. This situation is not uncommon. For example ntpd works out of the box, and the conf file need only be changed if, for instance, one installs a local central ntp server. Then all ntpd client conf files must be changed to point to the new server name/address. > In the wheezy box' inetd.conf, I found, from the install: > >> amanda stream tcp nowait backup /usr/lib/amanda/amandad amandad -auth=bsdtcp >> amdump amindexd amidxtaped > > and on one of the squeeze boxes: > >> amanda dgram udp wait backup /usr/sbin/tcpd /usr/lib/amanda/amandad >> -auth=bsd amdump amindexd amidxtaped > > I pasted this line to the inetd.conf on the wheezy box, and amanda started > working (amcheck did anyway). That's a significant difference in the inetd > configuration. I expect some postinstall whinage, but never a broken inetd > configuration. And I've never heard of configuring an amanda client. > > Does anyone have an explanation of this? Is amanda going to be broken when I > update the backup box to wheezy? Is it already broken? It would appear that the amanda developers/maintainers changed the default protocol from udp to tcp somewhere between the version you're running on Squeeze and the version that ships with Wheezy. Apparently they decided that reliable transmission became more important than lower latency, or that they could maintain low latency while using reliable transmission. This change may be due to the adoption of gigabit ethernet. TCP over GbE with modern hardware should have lower latency than UDP over fast ethernet. I'm not an amanda user and I've read nothing on this specific topic. Just making an educated guess. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/514f00a5.4060...@hardwarefreak.com
