Re: best way to keep web servers in sync
We use ssh -> cvs -> perl -> ssh -> perl for this.
User commits a change to cvs, the loginfo process execs a perl
script that itself ssh exec's a remote program to cd to the
appropriate dir and cvs update the module.
The advantage is that the live web tree only needs one user with
write access to the live web tree, we have complete logging and
we have reversion capabilities. The disadvantage is that the two
perl progs have to be suid to the user on the remote box and
that user has two way passwordless authentication.
I know that wasn't really your question . . . but it's what we
PWGroup - Saskatoon