neykov commented on this pull request.
> + }
+ }, null);
+ }
+
+ @Override
+ public Iterable<Location> listLocations() {
+ return ImmutableList.of(
+ new
LocationBuilder().id("localhost").description("localhost").scope(LocationScope.HOST).build());
+ }
+
+ @Override
+ public VagrantNode getNode(String id) {
+ // needed for
BaseComputeServiceLiveTest.testAScriptExecutionAfterBootWithBasicTemplate()
+ // waits for the thread updating the credentialStore to execute
+ try {
+ Thread.sleep(200);
That's a very ugly hack to work around the following scenario:
* a call to `client.submitScriptOnNode(node.getId(), AdminAccess.builder()...`
* which schedules
`persistNodeCredentials.ifAdminAccess(runScript).apply(node1);`
* followed immediately by `client.getNodeMetadata(node.getId());`
Usually calls to `getNodeMetadata` involve network IO and take some time to
complete. This gives time to the `persistNodeCredentials.ifAdminAccess` call to
execute. In the vagrant case though `getNodeMetadata` completes quickly, before
`persistNodeCredentials.ifAdminAccess` has the chance to execute. This results
in the old credentials returned with the metadata.
Couldn't see a clean way to fix this from the provider. Could change
`submitScriptOnNode` to return a future that gets resolved only after the task
listener completes, wdyt?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds-labs/pull/160
