why torch a burning building?
Small daily WTF:
We have this one simple event reporting system which is used barely few times a day total, no more. Some employees fill in some simple text fields and also day, month and year in freeform number fields, each number invidually in their own fields. Yes it should have proper UI but its not possible, not made by us etc. Now webform displaying that data stopped working lately. After spending some hours to track why SQL returns nothing, I find reason: 29.2.2012 (leap day) someone has input event as normal, but he mis-wrote to year-field 2011 instead of 2012. End results, that event is marked to non-existant day 29.2.2011 which causes sqlserver CONVERT-function to die fiery death. Why we have to use CONVERT function to parse real datetime field from text fields, thats wtf of its own... data fields including year, month and day are in same database table cell. Another "cant touch, not made by us".
What are chances someone inputs that exact wrong value for that exact day which causes application to break...
Work is making everyone get their mac certification or get fired out of the blue. Should be easily doable, but I have to pay test costs up front and I fucking hate everything Apple. Best surprise ever.
I'd hazard a guess that maybe 1/40 of the computers we get in are macs, and there's already a few guys who handle them. The situation is even more because I deal with the company's shit, not customers', and of course there are no macs. It was even specifically mentioned that those in my position are no exception.
HA HA~ right to work state
Actually i'd get one myself just for that if I didn't think they were overpriced
ib4 'that's not real programming'. Either way it's not graphical work.
Originally Posted by Loire
So please keep in mind for this post, I have an NDA that stops me giving details...
Every month we patch our (windows) servers. Its done by WSUS and its all good.. We roll out to test/dev first then a month later to prod.. The exception to the automatic patching is when we have a cluster as we have tried and failed several times to get windows to patch, fail over, patch and fail back.. We do cluster by hand at really stupid times in the morning..
So we have an SQL cluster running a production database, it has a test counterpart. Of the services that use this database we have one I shall call 'new critical application 1' (or nca1). This application has a test install in the test environemt and 7 production installs in geographically seperate locations (pretty much where the users are + 2 in the DC). nca1 has been developed at great expense to fulfill a critical business criteria. It has been in the dev/test environment for over 2 years and has been in production for about 3 months.
So I get an email from a developer yesterday. Apparantly once a month at 4am on a Thursday their application stops talking to the database. Fortunatlly it was actually Thursday when he sent this and I had done the patching that morning, where I had patched the SQL servers this application relies on.. It gets disconnected form the database because we bounce the database between nodes on the cluster and then doesnt reconnect..
This has been going on for 3 months in production and 2 years in development.. THE WHOLE TIME ITS BEEN IN DEVELOPMENT IT HAS BEEN DOING THIS AND NO ONE THOUGHT TO FIND OUT WHY.
Of cource now this means rather than them fixing their shitty software, we have to change our patching procedures to restart the services on all 7 machines after bouncing the SQL servers..
< Jolin> you're prety too LanaTorrinOriginally Posted by lubica
All I can say - :inhousetools:
Vulnerant omnes, ultima necat.