reweave/weave
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
5,666 Commits 7 Branches 85 Tags
5c19efae4804e90f3457be2160d0430df0df85eb
Commit Graph
148 Commits
Author SHA1 Message Date
Bryan Boreham 4e37f5ec42 Add explanations of various claim and allocation fields 2017-02-13 17:31:02 +00:00
Bryan Boreham 7075b5d9fd Extend pre-existing address reclaim to docker containers 2017-02-13 11:30:34 +00:00
Bryan Boreham 39ac1a0037 Collect existing weave addresses at start-up, to avoid a race when 
doing it from the script

Note we need to run weaver in the host pid namespace, so it can search
for existing container IPs. Also now we are firing off sub-processes
it's better to not be PID 1 so Linux doesn't expect us to reap zombies.
 2017-02-10 15:10:02 +00:00
Bryan Boreham 30d6639a98 Refactor: move peer name identifier to DB package 2017-01-11 13:05:25 +00:00
Bryan Boreham e42600632f Don't release dead container's IP addresses if not asked to check it is alive. 2016-11-14 19:53:54 +00:00
Martynas Pumputis 8569cbd4a3 Force transferring of IPAM ranges in a case of Shutdown 
If no peer for transfer is found, we transfer them to
mesh.UnknownPeerName to notify the ring update tracker.

Makes the awsvpc tracker to remove local entries from
the VPC route table when the last peer calls "weave reset".
 2016-06-10 11:51:31 +01:00
Adam Harrison adcb1018df Merge pull request #2345 from weaveworks/issues/2344-persist-more 
Fix ring persistence after free space update
 2016-06-09 16:37:57 +01:00
Martynas Pumputis 9fc7be83c8 Create a Tracker when starting the router 
The tracker's HandleUpdate method is called by the ring update routines.

In addition, we add a dummy overlay for AWSVPC to make `weave (status|
report)` to produce reasonable reports which includes AWS hosts
connection state.
 2016-06-08 16:44:54 +01:00
Bryan Boreham e2bee6e8c6 Persist ring after updates caused by reportFreeSpace() 2016-06-08 11:52:38 +01:00
Martynas Pumputis f50fbc002e Change ipam Universe from Range to CIDR 
There was no code to initialize it as anything else, and this lets us
bake in an assumption that it will always be a CIDR.
 2016-06-01 17:06:31 +01:00
Adam Harrison f02b78ed73 Merge branch '1.5' 
Closes #2282
 2016-05-26 16:16:54 +01:00
Bryan Boreham cda1084777 Check validity of IP subnets 
Be more consistent about this - check in the same way whether at startup
or during allocation.
 2016-05-26 14:31:17 +00:00
Martynas Pumputis 178438c470 Merge branch '1.5' 2016-05-10 17:36:56 +01:00
Martynas Pumputis 632694472c Merge pull request #2281 from weaveworks/issues/2275-request-order 
Ignore IPAM requests for space until we have a ring

Fixes #2275
 2016-05-10 17:36:26 +01:00
Martynas Pumputis 14e1145972 Merge branch '1.5' 2016-05-10 17:34:08 +01:00
Bryan Boreham 16b75f899e Move sleep on reset from IPAM Allocator to weave script 2016-05-10 16:03:08 +01:00
Bryan Boreham c3abe4f7b2 If we don't have a ring, just ignore a request for space. 2016-05-10 15:33:13 +01:00
Martynas Pumputis f9157acced Fix logging 2016-05-10 15:18:10 +01:00
Bryan Boreham de9597b6fa Merge branch '1.5' 2016-05-09 14:38:00 +01:00
Martynas Pumputis d698aefbfc Make Claim to block if the IPAM ring is not established 2016-05-09 12:48:41 +01:00
Matthias Radestock 32ba82efaf Merge branch '1.5' 2016-05-05 14:50:00 +01:00
Matthias Radestock b1aa52b3ed Merge pull request #2238 from weaveworks/missing-persist 
Persist the IPAM ring on `weave reset` and `weave rmpeer`
 2016-05-05 14:41:53 +01:00
Adam Harrison 34b95ee36f Merge branch '1.5' 2016-05-05 10:11:50 +01:00
Bryan Boreham 3252a25e00 Broadcast initial ring after persisting it, in case we die between those operations 2016-05-03 17:23:00 +01:00
Bryan Boreham 6847a3551d Persist the ring in Shutdown() and AdminTakeoverRanges(). 
Also make AdminTakeoverRanges cause pending ops to be retried.
 2016-05-03 17:23:00 +01:00
Bryan Boreham 8405e0759a Clarify the function of the periodic ticker. 
- it either retries for consensus or retries operations we would have tried previously.
There is no point retrying pending operations if we are waiting for consensus.
 2016-05-03 15:13:20 +01:00
Bryan Boreham 5ac8cc74d3 Refactor nested if-statements as switch 2016-05-03 13:52:36 +01:00
Bryan Boreham f6966548eb Delete persisted data when IPAM range changes 2016-05-03 13:20:04 +01:00
Bryan Boreham fb3ed9c1a6 Re-organise checking of persisted data so the failure paths bail out 
of the function first.
Also bail out if there is no persisted ring, since persisted 'owned'
data isn't meaningful in that state.
 2016-05-03 13:20:04 +01:00
Bryan Boreham 61fc5d56fb Refactor: move initialisation in the case that we don't use persisted data 
out from loadPersistedData().
 2016-05-03 13:20:04 +01:00
Bryan Boreham e96b68f0aa Don't use persisted data until we are sure it matches 2016-05-03 13:20:04 +01:00
Bryan Boreham fd7fcc762e Clarify the function of the periodic ticker. 
- it either retries for consensus or retries operations we would have tried previously.
There is no point retrying pending operations if we are waiting for consensus.
 2016-04-29 13:59:25 +01:00
Matthias Radestock 72e134605c refactor: move pure code outside actor 2016-04-26 20:02:47 +01:00
Matthias Radestock 7cc2a1fe38 refactor: more descriptive function name 2016-04-26 17:33:04 +01:00
Bryan Boreham 88a044f7f9 Merge pull request #2184 from weaveworks/2181-excessive-ring-frag 
prevent excessive ring fragmentation due to repeated space requests. Fixes #2181
 2016-04-25 11:54:12 +01:00
Martynas Pumputis 849edb1768 Add allocator.PeerGone 
The function is registered within the mesh gc and it is responsible
for removing gone nicknames (#2057).
 2016-04-22 17:29:27 +01:00
Bryan Boreham 5e4ab7f307 Plumb isContainer through Claim calls 2016-04-18 10:16:54 +01:00
Bryan Boreham 2b159b0c75 Pass checkAlive flag through as isContainer in Allocate 2016-04-18 10:16:54 +01:00
Bryan Boreham 984ba8a377 Extend 'owned' structure to store IsContainer flag 2016-04-18 10:13:09 +01:00
Bryan Boreham 01be916c37 Add test to expose problem deleting non-container IPs 
and check dead containers are removed
 2016-04-18 10:13:09 +01:00
Matthias Radestock 27632b118f prevent excessive ring fragmentation due to repeated space requests 
Merging a ring now indicates whether the ring has been updated. This
prevents excessive allocation retries, which in turn can result in
repeated space requests - i.e. if an allocation cannot proceed due to
lack of space - which in turn can result in excessive ring
fragmentation since many of the requests will result in the creation
of a new ring entries, due to the need to split ranges.

Fixes #2181.
 2016-04-17 12:40:18 +01:00
Adam Harrison a09986baa2 Rename consense -> prime 2016-04-14 12:46:03 +01:00
Bryan Boreham 61b0d0ba9e Merge pull request #2149 from weaveworks/stop-test-router 
stop ipam tests from leaking routers
 2016-04-11 13:34:21 +01:00
Matthias Radestock 695babf7ba make Allocator.Stop() idempotent and non-blocking 
So that tests don't have to be careful about ensuring that allocators
are stopped exactly once, or rely on the magic mesh.ChannelSize for
non-blocking behaviour.
 2016-04-11 12:32:01 +01:00
Bryan Boreham ecf1887e5b Initialize Paxos quorum to 0 until this node needs to request consensus. 
With a 0 quorum we will participate in voting, but will not actively
try to form a ring.
 2016-04-11 11:42:24 +01:00
Bryan Boreham c12a52d20e Better log and status messages for Paxos observer 2016-04-07 12:50:52 +01:00
Bryan Boreham 0c59de40e7 Refactor Paxos status call to remove type assertion 2016-04-07 12:12:08 +01:00
Adam Harrison 436dc01f52 Merge pull request #2070 from weaveworks/issues/1721-quorum-callback 
Use current target list to compute quorum
 2016-03-29 11:11:30 +01:00
Bryan Boreham 2b617633ae Compute quorum based on the dynamic number of target peers 2016-03-24 12:06:59 +00:00
Tom Wilkie 7a7867ae16 Fix spelling mistakes. 2016-03-23 16:09:43 +00:00