Amazon's CTO denies CNB coverage, 'Amazon's prime day down is not caused by DB transfers from Oracle'

By Mark Mathosian

Immediately after the start of Amazon Prime Day 2018 which started at noon on July 16, 2018 in Japan time, an access failure occurred about 1 hour after the server went down. Based on Amazon's internal document, CNBC in the United States reported that "server down has changed DB from Oracle to its own", but Amazon's CTO (Chief Technology Officer) Werner It is a situation that Mr. Vogels completely denies "It is a foolish Mislead guidance article."

CNBC published an article saying that on October 23, 2018 local time, Amazon Prime Day server down was due to not transferring the database from traditional Oracle to AWS Aurora and not preparing enough. What I was featured in this situation is that the amount of "savepoint" generated by individual transactions has become enormous and the processing speed of the entire system has declined abnormally.

The reason why the server went down on the first day of Amazon Prime Day is "the reason why we changed the DB from Oracle to our company" - GIGAZINE

One of the factors that caused major problems is that Oracle and Aurora have different ways of handling "savepoints". Although savepoint is an important DB tool for tracking or restoring individual transactions, an enormous amount of savepoints are created due to an extremely large number of orders that occurred on prime day, and the processing speed of the entire system abnormally decreases It is said that it is shown in the report that it has gone.

But the next day after the coverage, Mr. Vogels CTO at Amazon posted an objection to Twitter. To avoid getting stuck with the number of characters restrictions, I wrote a sentence in the image and challenged a long sentence.

Mr. Vogels refuted CNBC's article as "absurd and misleading" in a refutation. In addition, "AWS and Aurora have nothing to do with prime day server down, this problem was caused by a problem with Amazon Retail's software," revealing another reason I insist that CNBC's article is incorrect.

Mr. Vogels also said, "CNBC reporters read about the problem that occurred in only one of the 185 fulfillment centers around the world, and at that fulfillment center a small shipment "There is only a delay in the issue," CNBC's article insists that attention was focused on completely different issues.

Andy Pavlo, an associate professor of database technology at Carnegie Mellon University who is a complete third party, participated in the event. A reporter who wrote an article on CNBC seems to have sent Amazon internal documents to several experts including Mr. Pavlo and requested analysis of the content. At that time, replying that "It is not important that the handling of Oracle and PostgreSQL savepoints is not important," the reporter seems to have been furious.

Pavlo also replies to Vogels' tweets by replying that "(article) is insignificant content".

The interactions between Mr. Vogels and Mr. Pavlo are also touched in detail in the following Publickey article.

It is not caused by Amazon prime day server failure, Amazon switched from Oracle to Aurora DB. Amazon CTO denies CNBC coverage - Publickey

in Web Service, Posted by darkhorse_log