Thursday, February 21, 2008

anger management

best intentions notwithstanding, to date the render farm hasn't seen too much use. ergo, the best time to do some administrative housekeeping and all that.

sometime in the past months, the farm had been regenerated (the tech's term for a fresh install of various os'es). not surprisingly, the impossible man's idea of a regen involved taking some shortcuts and precious little by way of communication/consultation...

suffice to say that some whole steps were left out - and those are the things that i've been spending the last few weeks sporadically addressing (between editing and workstation troubleshooting).

in a nutshell, i made sure that each machine in the under-100-strong farm had the proper user and password; copied over the temporary database and set permissions on the folder structure; verified host lists for the whole farm system (license servers, job servers, render units); and did some render benchmarks to check if the system was usable.

so far so good.

in the latter regard, i also finally managed to inject a whole bunch of in-house code into the render system's submission script (after a call-for-help email to the software's programmer...).

the circumstances surrounding that bit of coding eureka was odd indeed. i had actually been sitting on the tip that the programmer had given for a week or two. until one day, as i came into work, an officemate (who had taken a personal mouse home the day before - without telling me) announced the need for a replacement or wouldn't be able to work. i inquired as to why that had been necessary, and the snappish reply that the mouse was "personal" after all took matters swiftly downhill...

long story short, replacement mouse was found, more ill-considered behavior and words were expressed, a heated mutual apology then transpired.

however, it takes me quite a while to descend from the plateau of heatedness to which i had levitated - oddly enough, to take my mind off the bad trip morning, i decided to take on the task of grafting the non-working code to the submission script.

lo and behold.

the script finally worked.

although i am happy to have worked that programming thorn out, the notion of having to fly completely off the handle and then redirecting the anger to solve the problem is not something i'm looking forward to applying at the next coding roadblock...

Monday, February 11, 2008