[CI] 6.x packaging tests failing on Fedora

[DaveCTurner]

The 6.x packaging tests are consistently failing on Fedora. The first build that failed here was 303 at 2017-11-13T18:22Z but there were other failures going back to 292 at 2017-11-10T02:18Z so it's hard to pinpoint a bad commit.

not ok 213 [UPGRADE] start old version
not ok 213 [UPGRADE] start old version
# (from function `wait_for_elasticsearch_status' in file /project/build/bats/utils/utils.bash, line 426,
# (from function `wait_for_elasticsearch_status' in file /project/build/bats/utils/utils.bash, line 426,
#  from function `start_elasticsearch_service' in file /project/build/bats/utils/utils.bash, line 312,
#  from function `start_elasticsearch_service' in file /project/build/bats/utils/utils.bash, line 312,
#  in test file /project/build/bats/tests/80_upgrade.bats, line 58)
#  in test file /project/build/bats/tests/80_upgrade.bats, line 58)
#   `start_elasticsearch_service' failed
#   `start_elasticsearch_service' failed
# Making sure elasticsearch is up...
# Making sure elasticsearch is up...
# --2017-11-13 19:04:32--  http://localhost:9200/_cluster/health
# Resolving localhost (localhost)... ::1, 127.0.0.1
# --2017-11-13 19:04:32--  http://localhost:9200/_cluster/health
# Resolving localhost (localhost)... ::1, 127.0.0.1
# Connecting to localhost (localhost)|::1|:9200... failed: Connection refused.
# Connecting to localhost (localhost)|127.0.0.1|:9200... failed: Connection refused.
# Connecting to localhost (localhost)|::1|:9200... failed: Connection refused.
# Connecting to localhost (localhost)|127.0.0.1|:9200... failed: Connection refused.
# Retrying.
# 
# --2017-11-13 19:04:33--  (try: 2)  http://localhost:9200/_cluster/health
# Connecting to localhost (localhost)|::1|:9200... failed: Connection refused.
# Retrying.
...
:qa:vagrant:vagrantFedora26#packagingTest FAILED
REPRODUCE WITH: gradle :qa:vagrant:vagrantFedora26#packagingTest -Dtests.seed=2321C67D5DCE7378 
:qa:vagrant:vagrantFedora26#packagingTest (Thread[main,5,main]) completed. Took 7 mins 2.216 secs.
:qa:vagrant:vagrantFedora26#halt (Thread[main,5,main]) started.
:qa:vagrant:vagrantFedora26#halt
Putting task artifact state for task ':qa:vagrant:vagrantFedora26#halt' into context took 0.0 secs.
Executing task ':qa:vagrant:vagrantFedora26#halt' (up-to-date check took 0.0 secs) due to:
  Task has not declared any outputs.
Starting process 'command 'vagrant''. Working directory: /var/lib/jenkins/workspace/elastic+elasticsearch+6.x+packaging-tests/qa/vagrant Command: vagrant halt fedora-26
Successfully started process 'command 'vagrant''
==> fedora-26: Removing cache buckets symlinks...
==> fedora-26: Attempting graceful shutdown of VM...
:qa:vagrant:vagrantFedora26#halt (Thread[main,5,main]) completed. Took 7.58 secs.

FAILURE: Build failed with an exception.

@tlrx you fixed one issue with this build in d4b7f7a, could you have a look at this too?

[DaveCTurner]

Possibly related to #27340 and thence to #25003.

[tlrx]

Failures from build 292 to build 302 inclusive are all related to the POI version fixed in d4b7f7a.

I suspect that for the 303 and 304 builds the vagrant box was so slow so it took more than 60 sec to check that elasticsearch correctly started. I increased the timeouts to 120 sec in Failures from build 292 to build 302 inclusive are all related to the POI version fixed in d4b7f7a and since the builds are green again I don't think we need to do more than that.