Closed Pe4enie closed 6 years ago
Hey @Pe4enie thanks for the report. The -e MAX_OLD_SPACE_SIZE is broken in the current release, so when you tried to increase it, it isn't working. Let me fix that ASAP.
Best Peter
@Pe4enie released 6.2.2 with the fix, if you set the memory for node, that should at least work. However it seems like something else is wrong, you shouldn't get it with only five runs for one URL so let us keep it open.
We will do very little during the holidays on the project and if you wanna do some investigation, you can follow the instructions https://www.sitespeed.io/documentation/sitespeed.io/developers/#debugging-with-chrome
Best Peter
Checked it some more and I think this is a bug in PageXray (I get the same thing with both Chrome & Firefox).
@Pe4enie PageXray gets stuck at the moment of the redirect chain:
that's the cause.
Hey @soulgalore. Man it was fast :) Thank you so much for quick answer/investigation. Unfortunately, I won't be able to try the fix until 2nd of January. If there will be any questions, I'll try to provide information needed
@Pe4enie np, it will not work with the current patch so that is good to wait until next year to test it :) The fix I did only make it possible to increase memory, we need to fix that bug in PageXray for the page to work, I'll roll out a fix the coming days.
Pushed 6.2.3 today that fixes the problem, so it should be ready when you get back @Pe4enie :)
Hey @soulgalore, first thing in the morning tested if everything okay :) And well, it is, works as a charm, thank you so much!
great, thanks for getting back @Pe4enie
Hi there,
was trying to update my configuration to latest major release (tried few different minor ones - 6.2.1, 6.1.3, 6.0.2) from the version I was using before - 5.6.5 and started constantly running into 'FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory'
After googling a little saw that people had similar issue only when they were running a lot of runs on thousands of urls. I currently get this error after 5 runs on 1 url consistently.
Here is a sample output:
And here is a script i'm running it with:
where no-login-urls.txt is file which currently consists of 5 urls which do not require logged in state. As you can see I even tried increasing node old space memory pool. it didn't help.
Here is default.json I use (I changed real values to dummy values):
I'm running this stuff on Ubuntu 16.04.3 LTS