zuul/nodepool
Code Issues Proposed changes

Add error handling when cleaning up resources

Browse Source 
It is possible for a single provider to have problems (eg: auth) which
blocks all other providers behind it. This causes a problem as leaked
instances could never be cleaned up in providers that are working.

Change-Id: I97d05815ae4812ba9e2881e1207ee901d269e655
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
This commit is contained in:
Paul Belanger
2019-05-28 21:02:35 -04:00
parent d0bb3d004d
commit e02b9d6097
1 changed files with 6 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
nodepool/launcher.py
View File

@@ -527,7 +527,12 @@ class CleanupWorker(BaseCleanupWorker):
for provider in self._nodepool.config.providers.values(): for provider in self._nodepool.config.providers.values():
manager = self._nodepool.getProviderManager(provider.name) manager = self._nodepool.getProviderManager(provider.name)
if manager: if manager:
manager.cleanupLeakedResources() try:
manager.cleanupLeakedResources()
except Exception:
self.log.exception(
"Failure during resource cleanup for provider %s",
provider.name)
def _cleanupMaxReadyAge(self): def _cleanupMaxReadyAge(self):
''' '''