customize driver polling interval in large Kubernetes nightwatch clusters #2193
Conversation
taitelman
changed the title
taitelman
changed the title
|
Thanks for your contribution. There is also another related PR here #2144. Maybe these 2 could be combined? Also it would be interesting to hear a bit more about your usage of Nightwatch. Would you consider contributing parts of your solution or maybe write a blog post? If you could add a few lines about your project here #2154 that will be much appreciated. |
|
Indeed my PR duplicates #2144 trying to merge them.
|
|
ok. I've pulled and enhanced #2144 into this PR. |
| @@ -90,6 +90,9 @@ class BaseWDServer { | |||
| this.pollStarted = false; | |||
| this.processCreatedTimeout = null; | |||
| this.processStatusCheckTimeout = null; | |||
| // for multi docker container increase max status polling retries and intervals. | |||
| this.maxStatusPollTries = settings.max_status_poll_tries || BaseWDServer.maxStatusPollTries(); | |||
There was a problem hiding this comment.
BaseWDServer.maxStatusPollTries is a static property, not a method and also I don't see maxStatusPollTries being converted to an instance property, it is still static: https://github.com/taitelman/nightwatch/blob/fixDriverPolling/lib/runner/wd-instances/base-wd-server.js#L60
There was a problem hiding this comment.
correct. fixing...
maxStatusPollTries is converted to instance property in line 94 and later we use it here
checkPingStatus() {
if (this.statusPingTries < this.maxStatusPollTries) {...
however, it is a good question if we want to expose a getter for that instance member. I didn't see any external usage for such case.
I may have done a poor member name choice by using this.maxStatusPollTries and BaseWDServer.maxStatusPollTries in the same code
|
I don't see any commits from #2144 and also there are some issues here so I'll go ahead and merge the initial PR. |
|
ok. let me check this again |
…to fixDriverPolling Conflicts: lib/runner/wd-instances/base-wd-server.js
|
This issue has been automatically marked as stale because it has not had any recent activity. |
with regards to #2192
in large container swarm we need longer timeout to see that web driver is up and responding.
added customized way to increase that timeout.
Actual stress tests showed that it might take the chrome driver few seconds (6) to responde for
localhost:9515/statusendpoint.also: docker host machine will need to increase max file handlers due to the new spawn mechanism of the driver (used to be handled inside Selenium Java process)