We need to ensure the process stops so the shutdown hook is very important.
But if the remote server hangs we need to hang as well so no timeout on
stop, we wait for the remote process blindy.
Romain Manni-Bucau
Twitter: @rmannibucau
Blog: http://rmannibucau.wordpress.com/
LinkedIn: http://fr.li
So you are saying remove the shutdown hook and anything that has to to
with process termination etc? I don't recall ever having a problem with
this since vm6 though.
Andy.
On 22/07/2014 07:29, Romain Manni-Bucau wrote:
Sorry but -1000
Best way to leak processes and kill a CI machine (we did
Ok so no need to push it on trunk then ;).
BTW we should get in touch with infra to get thread stack in such a case
instead of hacking it if we find time
Romain Manni-Bucau
Twitter: @rmannibucau
Blog: http://rmannibucau.wordpress.com/
LinkedIn: http://fr.linkedin.com/in/rmannibucau
Github: http
Yes, was more of an experiment to see if this test would go further. Can
only do it on the buildbot, as I can't replicate it here.
Andy.
On 22/07/2014 07:29, Romain Manni-Bucau wrote:
Sorry but -1000
Best way to leak processes and kill a CI machine (we did it)
It is good to hang in this case
