On Oct 26, 2005, at 3:45 PM, James Melin wrote:
We are using a procedure developed by Steve Wehr of IBM where you
have a
master /opt/IBM/WebSphere directory that is then shared read-only with
other Linux guests.
Apparently there has been little latency in the directory structure
being
refreshed on the read-only systems after a change has been made on the
owning r/w system. This has stopped happening, and I'm wondering
how you
make Linux refresh the directory of the r/o disk?
You could unmount it and remount it.
But you really shouldn't be doing this. The file system should be
read-only to EVERYONE. Otherwise it's perfectly possible for the
read-only guests to have very mutually inconsistent ideas of what's
in the filesystem: if it's a read-only fs, then there's no reason for
a guest to ever reread a disk block if it's still in cache, and so
huge changes can go by without your guests being aware of them. If
you actually need something that propagates updates in real-time, you
should use a network filesystem. At the VERY least, use a
journalled filesystem rather than straight ext2, so at least what the
r/o guest thinks is on the disk makes *some* sense.
Adam
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
