search

openshift-kni / eco-goinfra

Apache License 2.0
7 stars 33 forks source link

nodes and apiservers pkgs: reduce glog level from 100 to 80 #404

Closed elenagerman closed 4 months ago

elenagerman commented 4 months ago

the default is 100 for almost all our methods and it's very hard to debug on nodes reboots

elenagerman commented 4 months ago

@kononovn @trewest @achuzhoy WDYT?

kononovn commented 4 months ago

@elenagerman According to our original agreement all logs in eco-goinfra project should have level 100. If you want to reduce number of logs your could delete them.

kononovn commented 4 months ago

@elenagerman If you want to hide all logs from eco-goinfra you can set env var export ECO_VERBOSE_LEVEL below 100

elenagerman commented 4 months ago

@elenagerman If you want to hide all logs from eco-goinfra you can set env var export ECO_VERBOSE_LEVEL below 100

I didn't planed to hide all logs. The problem is that during long wait-for-something methods you getting such logs non stop and, on some stage losing relevant logs. As an example let's take WaitForAllNodesToReboot method: Wait for the reboot 1 node takes at least 5 mins. according to the pool size it will be biggest pool size * 5 mins. Each 2 second both Exists and IsReady print 1 message. Maybe we need to consider this option and to make some role related for glog??

kononovn commented 4 months ago

@elenagerman There are 2 options

  1. Redirect logs to file then you'll not loose any logs
  2. Delete these logs from functions.
  3. Open an issue and it'll be processed by maintainers.

Personally I would prefer first or third option.