[p4] Perforce Proxy and Checkpoint
steve at vance.com
Tue Feb 12 09:05:21 PST 2008
This is normal. If you had local users hitting the server during that
time, they would hang, too. The server locks all tables for
checkpointing. There's no way to change that.
The two most common solutions are variations on a theme. One is to use a
file server with the ability to take quick snapshots. Another is to copy
your journals over to another location, through real-time tailing,
frequent 'p4d -jj' invocation, or p4jrep. In both situations, you need
to start another server against the replicated/frozen data and do your
checkpoint and backup from there. It may require a secondary server
license (ask support) and there are some tricks to advance the journal
For NetApp filers, Richard Geiger has a script in his portion of the
Guest Depot. I have an updated version of it in my area that
parameterizes the mount hierarchy a little more. I used this to take a
client's lockout time from 45 minutes to 20 seconds. There are other
adaptations floating around for other filers and file systems, but this
one has the locking order and considerations correct and is a known good
Nellie Chen wrote:
> We have a team that is using a Perforce proxy server to access our central Perforce server. However, when we are performing a backup of the central Perforce server by doing checkpointing, the proxy server accessing the central server also hangs. Is this the expected behavior?
> The proxy server users have reported that they can neither view the files that are already synced to the proxy server, nor can they use p4 via command-line when setting P4PORT to the proxy server (p4 simply hangs).
> Is there a way to allow the proxy server users to access our central Server (just read access to the databases) while the checkpoint is occurring? If not, can you suggest a workaround?
> perforce-user mailing list - perforce-user at perforce.com
More information about the perforce-user