Hi Alexey
21.07.2016, 09:13, "Алексей Плотник" <
odis...@gmail.com>:
> I'm testing buckrunner daemon under the load. Backrunner proxies requests to 2 back-end nodes, 2 replicas. I'm using SSD.
>
> What I have tried is to ran 6 threads (terminals), every send 10,000 inserts of different objects of the same size. Simply I used curl to do that:
>
> curl -X POST
http://localhost:9090/upload/b1/object-w1-1 --data "20KB of data"
>
> "w1" is for worker 1, "w2" is for worker 2 and so on. One-by-one, I tried to upload 10K objects (x6 workers) = 60K objects
>
> Eventually backrunner process failed. Error is:
>
> runtime: garbage collector found invalid heap pointer *(0xb4cd60+0x240)=0x1 s=nil
> fatal error: invalid heap pointer
We will take a look at it.
Please tell us what were backrunner and elliptics-go commits used to build a server?
This info is provided in backrunner's /proxy_stat/ handler, please show it here.
Also what were the elliptics client version?
It is not allowed to change the library without recompiling backrunner and elliptics-go, otherwise really weird bugs occur.