Commit Graph
4305 Commits
Author SHA1 Message Date
Ilya Dmitrichenko f4ad97b0c9 Merge branch '1.4' 2016-02-22 14:16:37 +00:00
Ilya Dmitrichenko b377207faf Merge branch 'latest_release_doc_updates' into 1.4 2016-02-22 14:16:02 +00:00
Ilya Dmitrichenko 24ebb7cc7c Add missing dependency for latest gsutil 2016-02-22 14:11:36 +00:00
Ilya Dmitrichenko b5518db03c Add missing argument to tar 2016-02-22 14:02:39 +00:00
Ilya Dmitrichenko 2aa4e6a83a Use "official" gsutil installation instructions 2016-02-22 13:58:42 +00:00
Ilya Dmitrichenko 7c655cc3ba Fix false-positive docs uploads 2016-02-22 13:56:49 +00:00
Matthias Radestock d7a1bdd7a8 Merge branch '1.4' 2016-02-22 13:55:45 +00:00
Matthias Radestock 75cda01167 Merge branch 'latest_release_doc_updates' into 1.4 2016-02-22 13:55:18 +00:00
Matthias Radestock 4f0f3cf98d warn against using the plugin in combination with --net=weave
and expand the basic proxy usage docs slightly.
2016-02-22 13:53:39 +00:00
Matthias Radestock d9637ce620 Merge branch '1.4' 2016-02-22 13:38:08 +00:00
Matthias Radestock e014140134 Merge branch 'latest_release_doc_updates' into 1.4 2016-02-22 13:36:33 +00:00
Matthias Radestock c8ae128e59 for Travis, this no longer is a go project
Travis insists that one picks a project language. There is no go left
in the steps executed by travis. Instead we have Python, Ruby and
shell. Let's pick the first...

Picking Python makes travis unhappy with "Can not perform a '--user'
install. User site-packages are not visible in this
virtualenv". Dropping the '--user' flag fixes that.

(cherry picked from commit 1d8d098450)
2016-02-22 13:29:33 +00:00
Matthias Radestock 12d848162d shrink Travis usage
There is no point running builds, tests, lints on Travis, since we
already do that on Circle.

The only thing we aren't doing (and can't do) on Circle is doc
publishing, so now Travis does just that.

(cherry picked from commit 1038cbd1a3)
2016-02-22 13:29:33 +00:00
Matthias Radestock b12096bd6c tweak 2016-02-22 13:23:29 +00:00
Matthias Radestock 644485d4ae Merge branch '1.4' 2016-02-22 13:13:48 +00:00
Matthias Radestock 1430982166 Merge branch 'latest_release_doc_updates' into 1.4 2016-02-22 13:13:13 +00:00
Matthias Radestock 2a83aad1d6 warn against using plugin and proxy simultaneously
and give a complete example.

Partially addresses #1988.
2016-02-22 13:11:02 +00:00
Matthias Radestock 0f9d9c77a3 remove cruft 2016-02-19 00:30:22 +00:00
Matthias Radestock 33c81add8d update dependency to fix build breakage 2016-02-19 00:27:45 +00:00
Matthias Radestock 497a104db7 Merge branch '1.4' 2016-02-18 21:59:42 +00:00
Matthias Radestock 5e761a62b0 Merge pull request #1985 from weaveworks/issues/1984-discovery
Add two functions required by ipamapi "contract"

Fixes #1984
2016-02-18 21:55:53 +00:00
Bryan Boreham b442567406 Add two functions required by ipamapi "contract" but not actually used 2016-02-18 15:29:10 +00:00
Adam Harrison 5dd8584fb6 Fix veth cleanup on Ubuntu 15.10
Cope with the @ characters that have started appearing in veth names as
of Ubuntu 15.10.
2016-02-18 12:11:48 +00:00
Matthias Radestock 4bd18737b7 fix typo 2016-02-18 11:39:54 +00:00
Matthias Radestock 09cd9e091a Merge branch '1.4' 2016-02-16 14:09:06 +00:00
Matthias Radestock 17d53af2a2 improve logging of dns entry creation
- log inside the lock, so we get proper serialisation
- include the container name, so we can identify the specific entry
2016-02-16 14:04:34 +00:00
Adam Harrison 2d4d3577ba Merge pull request #1971 from weaveworks/issues/678-persist-ipam2
Persist IPAM ring
2016-02-15 10:15:53 +00:00
Bryan Boreham 546f9592f9 Hard-fail on persistence errors, otherwise the network can get confused on a restart.
Create `Allocator.fatalf()`, for consistency
2016-02-12 18:12:26 +00:00
Bryan Boreham 06f3793f05 Remove persisted IPAM ring data on peerName change 2016-02-12 18:12:26 +00:00
Bryan Boreham f396c60402 Persist the Ring 2016-02-12 18:12:26 +00:00
Bryan Boreham ad036499eb Add db infra to IPAM 2016-02-12 18:12:26 +00:00
Bryan Boreham 38c6eb28e8 Add basic infrastructure to use BoltDB for persistence 2016-02-12 18:12:26 +00:00
Matthias Radestock 11e2515f70 Merge pull request #1968 from weaveworks/refactor-veth-creation
"refactor" veth creation
2016-02-10 13:49:13 +00:00
Matthias Radestock c79e036c08 Merge pull request #1966 from weaveworks/issues/1965-fix-local-weave-ps
Make weaveutil container-addrs observe $PROCFS

Fixes #1965.
2016-02-09 22:40:36 +00:00
Matthias Radestock 7a19c2f147 add a test 2016-02-09 22:26:01 +00:00
Adam Harrison 4cee647e73 Make weaveutil container-addrs observe $PROCFS 2016-02-09 22:25:47 +00:00
Matthias Radestock 7ea3299dc6 oops
/usr/local/bin/weave: 1172: [: true false: unexpected operator

This got broken in #1941.
2016-02-09 22:05:11 +00:00
Matthias Radestock 5707581ba3 Merge pull request #1964 from weaveworks/weave-veth-cleanup
Fix attach-bridge bugs
2016-02-09 21:11:27 +00:00
Adam Harrison 264d5e4828 Make attach-bridge iptables rule insertion idempotent 2016-02-09 20:42:59 +00:00
Matthias Radestock 9311294d9c "refactor" veth creation
Invoke the new create_veth from a few places, eliminating code
duplication and inconsistencies.

Semantic differences:

- init_bridged_fastdp now skips the veth creation/config if the veth
  already exists. This eliminates routing blips. Though in reality
  something very strange must have happened to end up in this code
  with an existing veth.
- init_bridged_fastdp no longer deletes the $BRIDGE and $DATAPATH when
  veth creation/configuration fails. This is consistent with
  init_bridge, which does not delete $BRIDGE when something goes
  wrong. And it is generally desirable since the init_* functions are
  invoked on potentially existing bridges - e.g. following 'weave
  stop'.
- attach_bridge and setup_router_iface_bridge used to mark one veth
  end 'up', then invoke add_iface_* for it, and then mark the other
  end 'up'. create_veth marks both ends 'up' first. This makes no
  practical difference in these cases.

notable other differences:
- we detect veth presence by checking whether *both* ends are present,
  rather than just one. We deliberately continue to (and fail in) the
  veth creation if one end already exists. This can only happen when
  we've stumbled across a different interface, since a veth end cannot
  exist independently. So it is right to treat this as an error.
- init_bridged_fastdp invokes add_iface_bridge instead of duplicating
  what that does. Also makes for nice coherence with the
  add_iface_fastdp it is invoking too.
2016-02-09 19:22:14 +00:00
Adam Harrison 2bfc86e7b2 Cleanup attach-bridge iptables rule 2016-02-09 16:51:04 +00:00
Adam Harrison 36ee8bacb0 Remove all vethwe* interfaces in reset 2016-02-09 16:48:03 +00:00
Matthias Radestock 1bb1108b91 tiny refactor: use ${CONTAINER_IFNAME} for name construction 2016-02-09 16:40:58 +00:00
Matthias Radestock a77201bf77 Merge pull request #1955 from weaveworks/weave-attach-bridge
'attach-bridge' feature to link an existing bridge netdev with the weave bridge
2016-02-09 15:45:28 +00:00
Adam Harrison a3d842bf4a Implement attach-bridge 2016-02-09 15:20:11 +00:00
Adam Harrison 43eb70193b Merge branch '1.4' 2016-02-09 11:40:14 +00:00
Adam Harrison 1228dbddb9 Add release v1.4.4 v1.4.4 2016-02-09 11:20:51 +00:00
Adam Harrison a401d0d6ea Merge branch '1.4' 2016-02-09 10:53:52 +00:00
Adam Harrison c00b00d623 Merge pull request #1963 from weaveworks/issues/1959-dont-remove-network
Don't remove weave network on plugin shutdown because it bricks Docker 1.10
2016-02-09 10:53:13 +00:00
Bryan Boreham 7bbc9f1b82 Don't remove the weave network on shutdown because it bricks Docker 1.10 2016-02-09 10:02:06 +00:00