GitHub user neykov opened a pull request:
https://github.com/apache/incubator-brooklyn/pull/542
Mirroring entities should wait for a proper restart of remote
There is a window of time when restarting Brooklyn where it returns 404 for
its entities, making mirroring entities unmanage themselves. Fix it by
extending the existing filtering logic to return 403 for explicitly marked
endpoints while remote is STANDBY or rebinding.
I had to use a delay after master promotion because there is still a small
period of time just after the node becomes master but before it starts
rebinding where it will return 404 for entities and there are no flags to check
for it.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/neykov/incubator-brooklyn
brooklyn-node-restart
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-brooklyn/pull/542.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #542
----
commit bdd90d1f1e6e5573bad7bdb28d2f09438e98a2c3
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T10:52:29Z
Replace usage of STOP_MACHINE=true with STOP_MACHINE_MODE=IF_NOT_STOPPED
commit 332a7812ec6835cf3584dd0e3e03e1be17da8462
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T10:54:00Z
Add {pre,post}RestartCustom hooks
Can't rely on {pre,post}Stop/Start hooks being called on restart.
commit 668b88f1ec7ba2e8069968e7f9cfae4bef313a0c
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T10:54:21Z
Shutdown BrooklynNode gracefully on restart
commit beda5e62adb26f58e3e355bc230b65a508738296
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T11:19:48Z
Improve documentation for STOP effector parameters
commit 2ddecbe5bf06089536f459272c9863ad93d13949
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T15:08:35Z
Don't shutdown BrooklynNode if STOP_PROCESS_MODE says so.
commit 63af576254303ba792407c03bcc9aa6318ddcaa9
Author: Svetoslav Neykov <[email protected]>
Date: 2015-02-27T15:08:46Z
Umanage BrooklynNode only if machine is released.
Previous logic of looking at the effector paramaters wasn't working in all
cases as IF_NOT_STOPPED depends on external factors. Instead check if the
SshMachineLocation is still around.
commit 827a828d5bfd496a735e2fc7cb33181d5ca9783d
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-04T14:50:00Z
Decline requests for resources relying on rebinded state in non HOT states
No point in asking for applications when HA state is STANDBY. For example
404 response should mean that the resource doesn't exist, not that it's not
loaded yet.
commit 1242a73560a05f55234e359afd16b60498071f74
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-04T15:05:56Z
Deny resource requests during rebind
commit 0483bdd997f0b767584cc81689413623e3198713
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-05T13:06:20Z
Wait for the HA state to stabilize before letting requests in
There is a period of time when the state is MASTER, but the context is not
fully initialized.
commit d1880724cd806667355acb56917856be725e8ca9
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-05T13:07:53Z
Tests for the HA HOT state check for resource access
commit 0be15c5682b91ba7df23a9f1f136e76083bb07e8
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-05T13:11:07Z
Rename filter to follow naming convention
commit 09721506bdc7a0bb89a1e360c1b7ebb94cacfd44
Author: Svetoslav Neykov <[email protected]>
Date: 2015-03-06T12:02:31Z
Integration tests for resource accessibility with regard to the HA state
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
