Age | Commit message (Collapse) | Author | |
---|---|---|---|
2016-08-31 | removes hard coded paths and relies entirely on the following required ↵ | elijah | |
environment variables: LEAP_CMD TAG PROVIDERDIR LOGDIR LOCKDIR | |||
2016-07-26 | test cmd should use TAG | varac | |
2016-07-22 | Fix openvpn gateway ip | varac | |
2016-07-16 | Fix add_nodes | varac | |
2016-07-14 | remove verbose output | varac | |
2016-07-14 | filter creation of dynamic files | varac | |
2016-07-14 | improved init_deploy | GitLab Runner | |
2016-07-13 | remove update_platform | GitLab Runner | |
2016-07-05 | ignore net.ssh.authentication.agent eror | varac | |
2016-07-05 | fix platform and provider path | GitLab Runner | |
2016-07-05 | Refactor for local builds | varac | |
It's now possible to deploy locally. Large parts still need to get refactored. This is mostly working for these commands: - reset_deploy - simple_deploy | |||
2016-07-03 | remove trailing whitespaces | varac | |
2016-07-02 | fix puppet msg filter after cli->platform move | GitLab Runner | |
2016-06-16 | more robust checkout to switch from submodules to subrepos | varac | |
2016-03-08 | Fix dkim key msg | varac | |
2016-02-24 | ignore dkim msgs, dont run extra deploy on bootstrap_nodes | varac | |
2016-01-20 | extended puppet msg blacklist, make update_platform() failsafe | varac | |
2015-11-16 | use case-insensitive filter for puppet errors, checkout platform repo ↵ | varac | |
without much noise puppet 3 changed to uppercase 'Notice:' etc. | |||
2015-11-16 | ensure all nodes are started before deploying | varac | |
2015-10-31 | fix versions() call | varac | |
2015-10-31 | Print all errors after deploy run, add summary | varac | |
2015-10-22 | finally fixed exit status, don't duplicate errors on verbose mode | varac | |
2015-10-22 | minor fixes | varac | |
2015-10-22 | Improve Output, added verbose flag | root | |
2015-10-22 | fix exit codes on error | root | |
2015-09-13 | Merge branch 'master' of ↵ | varac | |
ssh://greyhound.leap.se:4422/usr/local/bin/platform_test | |||
2015-09-13 | updated msg filter to filter out flase-positive errors | varac | |
2015-09-11 | make sure vagrant.key has the right permissions. fixes #7447 | kwadronaut | |
2015-05-25 | force-run custom provisioning script, i.e. syncing the time before a deploy ↵ | varac | |
(#6935) | |||
2015-04-16 | use FAILURE_LOCKFILE for determining if deploy failure summary email needs ↵ | varac | |
to get send | |||
2015-04-15 | update_platform() is now able to updated even if branch was forcefully ↵ | varac | |
pushed, i.e. rebased | |||
2015-04-15 | Use two lockfiles now, one for indication a running process of ↵ | varac | |
leap-platform-test, and a different one that indicates the last deploy failed | |||
2015-04-15 | use /var/run/lock/ for lockfile, and use a branch specific lockfile name | varac | |
2015-04-15 | restructured the failure mails so they work correctly again | varac | |
2015-04-15 | fix lock test | Micah Anderson | |
2015-04-15 | lock file should be in /var/run | Micah Anderson | |
2015-04-14 | add a -l/--lock option to ci test script: | Micah Anderson | |
if this option is passed, and the lockfile exists, then deploys will not happen and an email will be generated. If it doesn't exist, a deploy will happen and then tests will be run. If the tests fail, the lockfile is set, and will not be removed unless tests pass, or if it is manually removed. I have set this option on the regular cron test, but not on the push to the repository, this makes it so a push to the repository will still allow deploys to happen, and removes the lockfile if tests succeed, otherwise the regular deploys will not happen (allowing us to investigate) | |||
2015-04-14 | reformat to a standard format | Micah Anderson | |
2015-04-14 | alphabetize flags | Micah Anderson | |
2015-04-09 | add an init_deploy, and reformat reset_deploy | Micah Anderson | |
2015-04-09 | move the failure emails into functions so they can be re-used | Micah Anderson | |
2015-04-09 | add destroy_deploy function that will remove the vagrant nodes (with | Micah Anderson | |
leap local destroy $vm), then run a bootstrap_nodes() on the nodes, which will do a leap local start on the node, then a leap node init, and then a leap deploy. This should do a full cycle test | |||
2015-02-11 | ignore updated leap_cli deploy messages | varac | |
2015-01-28 | moved timestamping to the right place | varac | |
2015-01-28 | added timestamps, really fail when there are other running procs | varac | |
2014-12-10 | show test option in the usage, don't check for running instances on test | varac | |
2014-12-09 | fixed minor bugs on create_provider() and add_nodes() | varac | |
2014-12-09 | run leap test with the --continue flag | varac | |
2014-12-07 | bootstrap: make sure machines are rebooted in order to be able to load ↵ | varac | |
kernel modules after a kernel update (#6494) | |||
2014-09-09 | fix bootstrap_nodes | varac | |