- Joined
- Sep 20, 2018
- Messages
- 4,144
- SL Rez
- 2004
- Joined SLU
- 2009
You'd think...but this large company is LL. Oopsies are what they do.I would think that a company as large as LL would not do this live.
Gee whillickers, who could have done such a thing?Someone really fucked up the SL forum.
As some one that does this for a living its not really that simple. There is always a test bed that is run where all this is tested out and make sure it runs correctly. But at some point you have to switch over to the production servers. This may sound like a simple task but it isn't always the case. Each system is different and there is always the chance that something will go wrong in a unexpected manner. This is why any company during changes overs calls these "outage windows" and not just "maintenance windows." No matter how well planned there is always a chance that something will stick its head up its ass and the whole thing will head south quickly.I would think that a company as large as LL would not do this live. It would be more like... Roll it to another server then point the domain there maybe?
We do a release 4 times a week, rolling over one webserver at a time out of rotation so end users never see a server in transition. The new software has been in staging, actively used by all employees, 24 hours before release. Taking down the whole site for an upgrade is so 2005.As some one that does this for a living its not really that simple. There is always a test bed that is run where all this is tested out and make sure it runs correctly. But at some point you have to switch over to the production servers. This may sound like a simple task but it isn't always the case. Each system is different and there is always the chance that something will go wrong in a unexpected manner. This is why any company during changes overs calls these "outage windows" and not just "maintenance windows." No matter how well planned there is always a chance that something will stick its head up its ass and the whole thing will head south quickly.
An in that time you have never had one thing go wrong and take the whole site down or severely impacted the service in some way? I'm not saying they took the whole site down for maintance. I'm saying "shit happens."We do a release 4 times a week, rolling over one webserver at a time out of rotation so end users never see a server in transition. The new software has been in staging, actively used by all employees, 24 hours before release. Taking down the whole site for an upgrade is so 2005.
Devotion to Duty
xkcd.com
I can't recall the last time we had to post an outage window or exposed a broken webserver to the customers.An in that time you have never had one thing go wrong and take the whole site down or severely impacted the service in some way? I'm not saying they took the whole site down for maintance. I'm saying "shit happens."
Hence the "someone" in the thread title.This is a managed hosted service by invision boards, they do the upgrades and such, all LL is doing is pointing their url to invisions hosted services.
They usually require help.Between this and the AWS outage the other day... I'm thinking every now and then a network engineer just snaps, and decides the internet was a bad idea, and tries to unplug it.