~hduran-8/juju-core/add_state_network_check_on_specified

~hduran-8/juju-core/add_state_network_check_on_specified_machine » Changes from revision 2585

From Revision 2585 to 2566

Rev	Summary	Authors	Date
2585	Ensured there is a retry when txnrevno of unit does not matc... Ensured there is a retry when txnrevno of unit does not match	Horacio Durán	10 years ago
2584	Go formatte Go formatte	Horacio Durán	10 years ago
2583	Go formatted Go formatted	Horacio Durán	10 years ago
2582	Refactored assignToMachine Added proper tests in separate fi... Refactored assignToMachine Added proper tests in separate file, added clone to unit and machine	Horacio Durán	10 years ago
2581	Go Formatted, added extra comments for context Go Formatted, added extra comments for context	Horacio Durán	10 years ago
2580	Changed the way networks are obtained. Changed the way networks are obtained. Fixed the corresponding tests Added a draft of check for changes	Horacio Durán	10 years ago
2579	Ran go fmt Ran go fmt	Horacio Durán	10 years ago
2578	Added rough check for network compatibility when adding to m... Added rough check for network compatibility when adding to machine	Horacio Durán	10 years ago
2577	[r=dimitern] environs: Add []NetworkInfo to StartInstance [r=dimitern] environs: Add []NetworkInfo to StartInstance This does not introduce new logic, just changes the StartInstance() return values (across all providers) to include a slice of NetworkInfo, which will be used in a follow-up by the MAAS provider to report what networks and network interfaces will be configured for the instance. https://codereview.appspot.com/84470044/ R=fwereade, gz	Dimiter Naydenov	10 years ago
2576	[r=dave-cheney] Always supply fake amd64/precise tools [r=dave-cheney] Always supply fake amd64/precise tools This is another place https://codereview.appspot.com/84280044/ needs to be applied https://codereview.appspot.com/84860046/	Dave Cheney	10 years ago
2575	[r=thumper] Add debug-log server api [r=thumper] Add debug-log server api Add the server side end-point for the debug-log api call. This is implemented as a websocket, with the request parameters defining the filtering behaviour, and the log lines that match being streamed back to the user over the web-socket. The first line over the websocket is the an error structure. If there are any issues either processing the parameters or opening the log file, there is an error, and if not, the Error component is 'null'. The capabilities of the server match the older python version as closely as possible. https://codereview.appspot.com/84880044/	Tim Penhey	10 years ago
2574	[r=hduran-8],[bug=1291022] Changed timeout on mongo wait aft... [r=hduran-8],[bug=1291022] Changed timeout on mongo wait after restore Bug 1291022 was caused by a race condition. After restore, there was a period of wait until mongodb would come up. The timeout value was set too low for cases where connection to the server was too fast. On slow connections there was extra time produced by the lag overhead. I changed the timeout value and tested in a machine in the same region than the environment created, it stopped failing after the change. https://codereview.appspot.com/82280045/ R=natefinch, rogpeppe R=natefinch, rogpeppe	Horacio Durán	10 years ago
2573	[r=wallyworld] Remove obsolete unit address methods [r=wallyworld] Remove obsolete unit address methods The SetPrivateAddress() and SetPublicAddress() methods on unit and uniter are no longer required. So this branch nukes them. It includes a change to the mega watcher tests because watching a unit no longer gives an event when the addresses change as these are now on the assigned machine not the unit. The unitDoc address fields can be removed when we support schema upgrades. https://codereview.appspot.com/82890045/	Ian Booth	10 years ago
2572	[r=dave-cheney] Always provide fake precise/amd64 tools [r=dave-cheney] Always provide fake precise/amd64 tools Fake tools take their template from version.Current, which gets weird when the machine running the tests is nor amd64 or precise. This is an attempt to fix this and improves many of the failing provider tests on ppc64el. https://codereview.appspot.com/84280044/	Dave Cheney	10 years ago
2571	[r=wallyworld],[bug=1302313] Cherry pick r2261 from 1.18 to ... [r=wallyworld],[bug=1302313] Cherry pick r2261 from 1.18 to allow uploading release versions.	Ian Booth	10 years ago
2570	[r=rogpeppe] state/apiserver: connection dies when State dea... [r=rogpeppe] state/apiserver: connection dies when State dead When the State connection has died (for example because of a Mongo leadership change or a server going down) we want to make it reconnect, so we make the Server die when that happens. https://codereview.appspot.com/84540044/	Roger Peppe	10 years ago
2569	[r=dimitern] api/provisioner: Allow adding networks and NICs [r=dimitern] api/provisioner: Allow adding networks and NICs This adds two new Provisioner API calls: * AddNetwork (exposed as st.AddNetworks bulk call on the client-side provisioner API) * AddNetworkInterface (exposed in the client-side API as machine.AddNetworkInterfaces bulk call) These will be used to set the networks/interfaces that will be configured by MAAS at provisioning time. Next, we'll make the necessary changes to StartInstance() in general (and in provider/maas specifically) to return what networks will the machine start with. https://codereview.appspot.com/84570043/ R=gz	Dimiter Naydenov	10 years ago
2568	[r=rogpeppe] worker/singular: mongo sanity-check tests [r=rogpeppe] worker/singular: mongo sanity-check tests All the previous tests of this package test that the logic works as written, but nothing checks that the approach will actually work when integrated with mongo. These tests actually start a replica set and mimic some of the structure that we intend to use in a full juju installation. By their nature, the tests are slow (~ 2 minutes to run) and unreliable (mongo occasionally core dumps), so I have disabled them by default. They can be enabled with the -juju.unreliabletests flag. https://codereview.appspot.com/84470043/	Roger Peppe	10 years ago
2567	[r=dimitern] state: Add machine networks and interfaces [r=dimitern] state: Add machine networks and interfaces Added a couple of entities (and related methods) in state: * Network - holds a name, CIDR and VLAN tag; used to describe actual configured networks that machines need to be on. * NetworkInterface - linked to a machine id and network name, holds the network interface name to MAC address mapping in state (MAC address as key for uniqueness); machines support adding one or more NICs. Renamed machine.Network() and st.networks to RequestedNetworks() and st.requestedNetworks (in code only, schema remains the same). This helps avoid confusion between the requested networks (to include/exclude on a machine, coming from the service; they are now called "requested networks"), and actually configured networks on a running machine (just "networks"). For now networks are immutable, can't be removed and NICs can be added to them only if their machine is not provisioned yet. NICs are only removed when the machine itself is removed. Provisioner API change: renamed Networks() to RequestedNetworks() which is fine, because Networks() was added recently and not released yet. We'll use the added machine NICs to show enabled networks in juju status later on, and they will be set at StartInstance() time inside the provisioner. Internally, MaaS provider changes will be needed to get NIC-to-MAC mapping for an acquired node and the list of networks for it. https://codereview.appspot.com/83070047/ R=fwereade	Dimiter Naydenov	10 years ago
2566	[r=jameinel],[bug=1290843] state/apiserver: rate limit Login... [r=jameinel],[bug=1290843] state/apiserver: rate limit Login requests As part of scale testing, I noticed that when you have 1000s of units and you bounce the API Server, everything comes flooding back. It appears that all of the Login requests end up queued in Mongo, to the point that the system hangs. This changes the Login request so that it limits the maximum number of concurrent requests (current value of 10 concurrent requests). It uses a new helper in utils/limiter.go that just creates some nice structure around a channel. We could probably structure it differently, or just use the channel directly, but it made it easier for me to understand what was going on. I also set it up so that requests for a User to Login are not rate limited. That way even when restarting and all the agents are hammering away at the server, you should still be able to run things like "juju status". I played around with the actual concurrency a bit, but 10 seemed to work just fine. The time for the system to come back to stable wasn't particularly affected (when they are going smoothly, Login only takes <100ms, the problem is that when they were flooding, they could take >5s). https://codereview.appspot.com/83830045/	John Arbash Meinel	10 years ago

Older »