Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to email@example.com.
Alexey Panchenko wrote: > Probably this is caused by locking. Are you using lockserver? > > You can connect to it using any telnet client and execute commands: > - Monitor > - Locks > to list active locks. I'm not sure those commands work any more, but it's worth a try - monitor slows things down a lot though. > Each new lock request is iterating over all previous locks linearly > and if you have many files and many simultaneous users - that could be > the bottleneck. You'd have to have a *lot* of locks before that would be significant, and you only get those kind of numbers on large numbers of commits (which have to wait for each other anyway). For an update/checkout you typically have one lock per user active at any one time. I'd suspect something happening at the network layer - the machine isn't under load at all so the process isn't really doing anything. Might be worth uninstalling the AV as a test (merely disabling AV normally doesn't work due to the way they interact with the system). Also any onboard firewalls. Tony