[p4] Is a 1500 line protect table excessive?

[Brad Holt]

Honestly one of the best investments in time I've made at work was a project to tackle the protections and get them simplified.  I couldn't really speak to the performance gains you might see, but the task has saved me countless hours down the road.  After implementing a whole group/subgroup hierarchy, and rewriting the protections, just adding a new user has gone from being a bit of a chore to being absolutely trivial.  Another advantage to the smarter use of groups is that I now have users organized in p4 relative to their geo's as well.  So I can script up automations to ensure that a user in location x knows to use their appropriate proxy, or has compression set as it should.  Also in my absence now, it is easy for whomever is filling in to be able to add a new user.

There are no longer any distinct users in the protect table, just groups.  There almost no exclusion lines in the table, which makes reading much easier and gets rid of all those duplicated lines to precisely define what can be done.  I also no longer honor requests for all sorts of intricate special casing.  If something along these lines really must be done, then it's time for a trigger.

I'd highly recommend a cleanup.  I've just now inherited another couple of p4 servers to administer and their protection tables are all over the place.  I can't make any sense of it.  Horrible.

-----Original Message-----
From: perforce-user-bounces at perforce.com [mailto:perforce-user-bounces at perforce.com] On Behalf Of Ivey, William
Sent: Tuesday, December 11, 2007 1:49 PM
To: Perforce User
Subject: [p4] Is a 1500 line protect table excessive?

I'd think so.



I'm looking at the protect table of a server I wasn't the admin of and

it is right around 1500 lines (all groups). Looks like the only

wildcarded path sections are the usual trailing /...



Some of the lines are probably obsolete and a lot of them can

be replaced by making some groups sub-groups of others which

would replace three protect lines by two at least.



Does anyone have a feel for the performance impact this might

have if we could cut it down by 30-40%? (Is it worth doing a major

overhaul in the short term vs. whittling it down over time?)



Thanks.



-Wm







_______________________________________________
perforce-user mailing list  -  perforce-user at perforce.com
http://maillist.perforce.com/mailman/listinfo/perforce-user