I use yandex-tank to generate a load 1 post per 10 seconds during 24h. But yandex-tank failed after about 16 hours of running, because tank.log consumed all the free space available. In my case it increased till 37Gb.
My load.ini:
[phantom]
address=192.168.254.201 ;Target's address
port=12224 ;target's port
rps_schedule=const(0.1, 24h) ;load scheme
connection_test=0
ssl=0
My ammo.txt consists of 10 similar post requests:
300
POST /api/< maybe confidential data>dimension1,dimension2,channel HTTP/1.1
Host: <confidential data>:12224
Content-Type: application/json
Content-Length: 103
Connection: keep-alive
{ "dimension1":"dimension1_1", "dimension2":"dimension2_not_used", "channel":"channel_1", "value": 91}
The command line:
yandex-tank ammo.txt
It seems that space consumed by repeating records "Stats cache timestamps:", like
2016-12-28 04:52:21,033 [DEBUG] yandextank.plugins.Aggregator.plugin plugin.py:101 Stats cache timestamps:
[1482836903, 1482836904, ....]
in the beginning of the file, this record consists of 1 timestamps. But the last available "Stats cache timestamps:" contains 54212 timestamps! Ther are more than 3 billions of timestamps in the file totally!
Is there a way to suppress/switch off this logging?
It's a bug. I've removed these messages.