View Profile
View Forum Posts
Private Message
Visit Homepage
Ahhh...sounds like you've outgrown that place!Originally posted by Venus Flytrap+Nov 6 2004, 11:43 PM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (Venus Flytrap @ Nov 6 2004, 11:43 PM)</td></tr><tr><td id='QUOTE'> <!--QuoteBegin-SaintHax@Nov 6 2004, 08:18 PM
I like what I do but I don't like the way it's done...put it this way, if we cut corners anymore than what we already do there wouldn't be anything left to cut[/b][/quote]
this place is full of BS
example:
we decided to offer Shoutcast stream services...ok to start...boss asked about it on our customer forum, I ended up answering him on our customer forum...I knew something was up...at this point no one had a clue we were gonna offer it...about 2 weeks later engineering gets an email saying set this up with a spec list...he already sold 900 listener slots to some online radio station before we even had a server online to run it...well I know shoutcast so I ended up taking that project on that night...
I can fireup a shoutcast stream in like 2 mins but when you do multiple stream services you have to sort that crap neatly, then if the server crashes it all has to automatically restart all the services on boot time(easy)
so a few hours into that and I had the linux server up with 6 service streams setup in their own directory each with their own config files and log files...then I set them up to load at startup should the server need a reboot. Everything was setup to be automated and trackable, it was a neat and clean setup...like I normally do things
now all in all that project went well as I didn't run into any major issues getting it built up...but wait it gets better...
a few days later I check the network graphs for the server and it's pumping out 50Mbits at peak load, I checked the listener level and there were only about 450 active streams, plus the 6 incoming source streams...this means when figuring in TCP overhead chatter the server will soon max out it's 100Mbit connection at about 93-95Mbits and require another network card...if it wasn't short notice to begin with then I would have built the server myself and packed it full of network cards.
that's just one example of how things are done backwards there, we had a router failure the other night, and it was the "brain" card of our #4 router that's in a data center across the street...it was midnight when it died...well the network acted up just a bit while the other routers canceled out the problem router and took over the traffic load...my boss had to come down and replace the card with me because no one ever showed me where the damn router was across the street. we rent a cabinet over there to plug that router into THE fiber hub for the Southeast then feed the bandwidth under the street into our data centers where the other two main routers are...and good luck finding that router in a 23000 sq. ft room packed full of network gear and servers.
So anyway that didn't seem too bad till we were over there with laptops loading the network settings on the new router “brain”, then he tells me "it's a good thing this card came in today" I thought "oh shit we didn't have this spare on site!?!?" so anyway that router was acting up a little the week before because the internal memory was bad and wouldn't let us write the config changes to it. So he ordered a new card and damn if it didn’t make in to the office within 12 hours of the bad part failing.
Now technically we could lose 2 of the 3 main routers and still run all the network traffic but that’s living on the edge in my book.
Reply With Quote