On Tue, Dec 23, 2008 at 10:39 PM, bofh <goodb...@gmail.com> wrote: > On Tue, Dec 23, 2008 at 10:48 PM, Andrew Konkol <akon...@gmail.com> wrote: > > > > since when does fortune _ _ _ _ _ have policy? > > Now now, I didn't say they were good policies. There's this wonderful > story I like to tell about my last place. You may have heard me > ranting about the NotWork Engineer there. But, this is a story about > their change management policies. > > We had a work at home project, and the VP told me to do testing in a > way that doesn't affect production. The said NotWork Engineer bought > a 100mb line that was doing *NOTHING* and had been doing nothing for 6 > months. Literally, bits were falling off the RJ45 onto the floor. So > I told the VP we can use that. > > The fucking fucker came up with a "top secret emergency CIO project > that needs to use the 100mb line". We currently have two BGP'ed 10mb > line for redundancy. > > Hold up pointer and middle finger, and twist them together. I showed > that to the change management folks. Told them that's current state. > Stuck thumb out, and told them that was the 100mb line. Told them > that the NotWork guy wanted to break the 2 10mb line. Untwist > fingers. 3 fingers are up right now. > > Told them that NotWork guy wanted to give one 10mb line to me for > testing. Fold pointer finger down. > > Told them Notwork guy wanted to use BGP to bond the 10mb (middle > finger) and 100mb (thumb). > > Told them he doesn't understand BGP and have no hope in hell of making > any of that work. > > Therefore, the 100mb line won't be used. > > Fold thumb away. > > Waved Figure 1 in their collective faces. Asked them if they > understood what that means if they allow the change to go through. > > They allowed the change to go through. I was right - the 100mb line > continued to stay unused. > > In their defense, the NotWork guy must have had photos of the CIO - he > doesn't do any work, and have caused a lot of outages (who the fuck > designs a 10/8 network?) and I once saw the CIO apologizing to him > after making a business decision he didnt like. For over 15 minutes. > "I want you to understand that it was nothing personal, blah blah > blah" > > Now that the bile has gone down some, I love telling this story, with > the accompanying finger gestures. > > > Seriously, production is a joke. > > So, yes, a lot of times. >
WTF are you talking about? You had me holding my hand up for a minute... Anyway, We use RT. We must be doing something right because we wrangle ~12K compute nodes with 3 people... Whatever system you choose is going to be useless unless you have management backing. Integration with AD or nis or whatever network service is key as well since without it , it will be yet another system for people to have a separate login. As far as home grown stuff goes, we have some django portal that is in house riddled with bugs and of course, the person who pushed it through is not with the company anymore. It's turned into a dinosaur that some people still rely on. To make it openbsd related, most people I work with have HEARD of OpenBSD but that is about it. the majority of stuff I see working for big oil is almost all commercial stuff. Pretty pink blue, red apliance boxes, rhel5u64 on the cluster nodes yah yah. As a matter of fact, the only BSD I see is the panasas sans which run a hacked up version of FreeBSD 4.6. Merry Christmas.
