Hi,
I' m using csync2 to keep a cluster of web servers synchronized. I've set
up csync2 in a chained
replication, meaning that create, update and delete events will go from one
server to the next
until the circle has been completed. This works well.
However, I've got one concern:
Let's say the cluster contains 4 web servers. The file 'index.php' gets
deleted from web01, and
the event gets propagated to web02, web03, web04 and finally back to web01
to complete the circle.
In this case, web01 won't do anything as the file has already been deleted
from this web server.
However, let's say that once 'index.php' has been deleted from web01, a new
version of 'index.php'
gets uploaded to web01. As far as I can tell, this leads to the newly
uploaded file being deleted
as well due to the fact that when the event comes back to web01 - then
'index.php' still appears
to exist. It is indeed a file with the same name...but in reality it isn't
the same file.
My assumption is that file comparisons are name based. However, is this
correct? Does csync2 use
others means to check if two files indeed are the same (besides from
checking path and filename)?
--
Stian Brattland
Gåsebæksvej 15, 1
2500 Valby
Danmark
_______________________________________________
Csync2 mailing list
Csync2@lists.linbit.com
http://lists.linbit.com/mailman/listinfo/csync2
