-
Hi! I was recently investigating a bug on production and discovered a weird behavior with JedisPool and AWS Elasticache. Basically, when Elasticache doesn't have enough memory left, the next write requests from Jedis are just... canceled. I cannot find another word for it :) No errors or anything, the program just continues to execute. Our code looks like this:
What happens is that during the iteration the Redis database becomes full, the iteration just stops without errors, and the code continues to swap the databases with only partially applied changes. Has anybody met this behavior? Is it something I have to tweak in Jedis or rather in AWS Elasticache? I tried to find similar issues but without luck. Any clue is appreciated. Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
@kalimatas Did you check the |
Beta Was this translation helpful? Give feedback.
-
Hi, Thanks for the hint: indeed, there are failed responses which we can extract and check. Here how it works:
|
Beta Was this translation helpful? Give feedback.
Hi,
Thanks for the hint: indeed, there are failed responses which we can extract and check. Here how it works: