An FYI but redis is absolutely affected by the recent meltdown/spector CPU design faults; https://gist.github.com/antirez/9e716670f76133ec81cb24036f86ee95
Patched hypervisor, unpatched guest:
Patched hypervisor, patched guest:
This add-on (and my redis add-ons) use pipelining when dealing with multiple redis instructions, and use lua to reduce the required redis instructions where possible.
But I don't think anyone is going to care about dropping from 159261 requests/s to 119502 requests/s (or ~30% hit) , the new redis connection per request will be a bigger issue
Patched hypervisor, unpatched guest:
Code:
Pipelined:
SET: 602772.75 requests per second
GET: 1012145.75 requests per second
Not pipelined:
SET: 170590.25 requests per second
GET: 159261.03 requests per second
Patched hypervisor, patched guest:
Code:
Pipelined:
SET: 554631.19 requests per second
GET: 857632.94 requests per second
Not pipelined:
SET: 121285.63 requests per second
GET: 119502.87 requests per second
This add-on (and my redis add-ons) use pipelining when dealing with multiple redis instructions, and use lua to reduce the required redis instructions where possible.
But I don't think anyone is going to care about dropping from 159261 requests/s to 119502 requests/s (or ~30% hit) , the new redis connection per request will be a bigger issue
It also dramatically reduces the amount of traffic the master-redis instance needs to handle.Ended up having this issue pop back up, created local read only slaves and the issue is gone. I think there may have been some latency created by the large number of requests to that single instance. Has been error free for a few days now!