by dklsf on 4/16/17, 3:08 AM with 5 comments
by atonse on 4/16/17, 12:07 PM
5500 is an odd number (not power of two) which rules out some kind of overflow issue.
Was this some kind of fixed size database? Why would it have crashed when storing 5500 rows?
Maybe storing that many hit a timeout, and maybe they were stored outside a transaction, hence leaving the system in an inconsistent state? But that wouldn't explain the database shutting down.
Maybe they were trying to store 5500 but hit a limit at 4096 before getting there?
I feel something is lost in translation (between nerds and non nerds) in the explanation for this outage.
by cnvogel on 4/16/17, 2:53 PM
http://www.raytheon.com/capabilities/products/autotrac/ https://www.youtube.com/watch?v=QB4Qy6mNYWY
by lm2s on 4/16/17, 9:51 AM
by cpncrunch on 4/16/17, 1:39 PM