From: <bac...@su...> - 2008-03-11 03:11:08
|
Searching through the mail archives, I see lots of posts about getting aborts with signal=PIPE on backups. I've got this problem on restore. I have been backing up 5 machines (linux and windows) using rsyncd flawlessly for a few months now. Thought it was about time to check restoring (backups are useless unless they restore!). I had no luck. I had two types of errors. The first thing I tried was restoring from the backuppc server (running debian testing/lenny) to a ubuntu 7.10 box. There, I got a failure with the signal=PIPE error. Responses to questions people have about this error during backup say that some big files can cause it. So I retried restoring a single small file and got the same result. I figured I would try restoring a single small file to other hosts on the network. I tried another debian box and a winxp box (both of which have been happily backed up for months). Both of those fail with "unable to read 4 bytes". According to Les Mikesell in a response (to someone having this problem on a backup), it means: > Usually this means that ssh did not authenticate correctly to start the > connection. Be sure you have tested the passwordless access running as > the backuppc user on the server (the key setup is per-user). Not sure where to go with that. I have rsyncd daemons running on all the hosts being backed up. According to step 5 of the docs, the rsyncd approach doesn't use ssh. Do I need to set up a passwordless ssh setup? I thought I have read through the docs pretty thoroughly but I haven't seen how to do this. I'll check again, but if someone can point me in the right direction, that would be helpful. Is the lack of a passwordless ssh also the cause of the signal=PIPE I see going to the ubuntu machine? Thanks. J.S. |