reweave/weave
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
2,130 Commits 7 Branches 85 Tags
dc8a82e9138f22bf81027ebdc2ed3ec27bdf23c6
Commit Graph
2130 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Matthias Radestock dc8a82e913 cosmetic: consistency in launchers 2015-06-03 14:22:20 +01:00
Bryan Boreham c1f8b62e11 Merge pull request #832 from rade/412_min_backoff 
LGTM. Fixes #412
 2015-06-03 12:47:07 +01:00
Paul Bellamy c91e699a27 Merge pull request #831 from rade/192_kill_container_on_network_config_error 
kill container on network config error
 2015-06-03 12:45:23 +01:00
Adam Harrison 4a8850afb3 Fix comment 2015-06-03 12:08:13 +01:00
Matthias Radestock d1c41a97b1 encapsulate retry delay calculation 2015-06-03 12:04:33 +01:00
Matthias Radestock fe5102580d fix connection retry delay calculation logic 
The delay d(n) for the nth retry was
d(1) = rand(initial_interval)
d(n+1) = d(n) + rand(d(n))

There are two problem with this:

1) The entire sequence is dominated by the random value for d(1). If
this is a small value, lots of retries at short interval ensue. In
extremis, if d(1) is 0, indefinite immediate retries would ensue.

2) The delay is not a true exponential backoff; the upper bound only
increases exponentially *probabilistically*.

The new algorithm calculates

d(n) = random_value in range [i-i/2,i+i/2],
       where i = initial_interval * 1.5^(n-1).

Thus d(1) is at least initial_interval/2, and successive values are
independent of previous values, forming a true exponential sequence of
intervals bounds between which the retry delay is chosen randomly.

Due to the change in algorithm we reduce...

- the initial interval from 5s to 2s. Recall that previously d(1) was
  in the range [0,initial_interval=5s]. It is now [2-2/2=1s,2+2/2=3s].

- the max interval from 10 minutes to 6. Previously, once that was
  reached, it would stay exactly there. It is now randomly chosen in
  the range [6-6/2=3m,6+6/2=9min].

Thus previously delays would range from 0 to 10 minutes, now they
range from 1s to 9 minutes.

Fixes #412.
 2015-06-03 11:29:55 +01:00
Matthias Radestock d909a86fb0 Merge pull request #828 from rade/666_no_timeout_on_wait_for_status 
no timeout on wait for status

Fixes #666.
 2015-06-03 11:20:37 +01:00
Matthias Radestock 4e68bde7b5 don't time out waiting for weave infrastructure containers to start 
instead keep trying, until we succeed or the container dies

Fixes #666.
 2015-06-03 11:19:51 +01:00
Matthias Radestock 52f31f0596 Merge pull request #830 from rade/829_fix_container_death_check 
kill -0 $CONTAINER_PID does not work inside weaveexec container

Fixes #829.
 2015-06-03 11:17:23 +01:00
Matthias Radestock fa8b77b060 kill -0 $CONTAINER_PID does not work inside weaveexec container 
since $CONTAINER_PID is in a different process namespace.

Fixes #829.
 2015-06-03 11:16:08 +01:00
Matthias Radestock 83fe4e59f6 kill container on network config error 
but only if we have just started it; killing other containers, e.g. on
`weave attach`, would be rather impolite.

Closes #192.
 2015-06-02 20:26:20 +01:00
Matthias Radestock 8de1ddd255 cosmetic 2015-06-02 20:21:56 +01:00
Matthias Radestock 395aea28c6 cosmetic 2015-06-02 20:16:25 +01:00
Matthias Radestock 5e7d1cd0eb Merge remote-tracking branch 'upstream/master' 2015-06-02 18:46:30 +01:00
Michael Bridgen 3344e45a9f Merge pull request #825 from squaremo/discover_ipam_ip 
Allow lookup of IPAM-allocated addresses
 2015-06-02 18:04:11 +01:00
Matthias Radestock d020999e82 cosmetic: more sensible function order 2015-06-02 17:52:39 +01:00
Matthias Radestock d105d255ca Merge pull request #824 from paulbellamy/weavewait-race-condition 
[proxy] We can only rely on receiving a signal when we are a container

This got broken in #820.
 2015-06-02 17:39:28 +01:00
Michael Bridgen 791584d9f3 Allow lookup of IPAM-allocated addresses 2015-06-02 17:31:25 +01:00
Matthias Radestock 48b54ef722 cosmetic: consistency 2015-06-02 17:25:56 +01:00
Paul Bellamy e2447721eb [proxy] We can only rely on receiving a signal when we are a container 
Due to https://github.com/docker/docker/issues/9098, there is no way to
signal an exec. Instead we'll have to just poll the interface and hope.

Adds `weavewait -s` which is the way to tell weavewait to skip waiting
for a signal, which we use on execs
 2015-06-02 17:21:41 +01:00
Bryan Boreham 89074715bc Merge pull request #823 from bboreham/818-claim-no-trunc 
Fixes #818
 2015-06-02 16:12:47 +01:00
Bryan Boreham 4968f0f3c4 Merge pull request #822 from bboreham/802-test-race 
Fix race condition in allocator fuzz-test - fixes #802
 2015-06-02 16:12:22 +01:00
Bryan Boreham aa7c7a9a9d Add --no-trunc flag to docker ps for ipam_claim 2015-06-02 16:07:48 +01:00
Bryan Boreham 6a1b658f59 Fix race condition in allocator fuzz-test 
It was possible to free an address between deciding to re-test it
and actually re-testing it.  So introduced a flag 'block' to prevent
an address from being freed during the re-test.
 2015-06-02 16:04:35 +01:00
Matthias Radestock e2805c01f4 cosmetic: remove stray trailing whitespace 2015-06-02 15:57:50 +01:00
Matthias Radestock 04c10233f7 Merge pull request #820 from paulbellamy/weavewait-race-condition 
Use USR2 signal to tell weavewait when to continue

Fixes #812.
 2015-06-02 15:52:22 +01:00
Paul Bellamy 92444a75eb [proxy] Use USR2 signal to tell weavewait when to continue 
There was a race condition with containers running and exiting before
`weave attach` was finished, which would cause an error.
 2015-06-02 15:48:49 +01:00
Paul Bellamy cb92e6b0d2 Merge pull request #817 from rade/420_move_exe_image_to_subdir 
move exe and docker image producing dirs to subdir
 2015-06-02 14:18:49 +01:00
Matthias Radestock 5b04efac7c move weaver 2015-06-02 14:04:02 +01:00
Matthias Radestock 485db6e48b move weavedns 2015-06-02 13:51:19 +01:00
Matthias Radestock ee28cf91e1 move weaveexec 2015-06-02 13:46:02 +01:00
Matthias Radestock 8d4a97fed5 move sigproxy, weaveproxy, weavewait 2015-06-02 13:19:16 +01:00
Matthias Radestock 46ef954057 cosmetic: rename: trains station 2015-06-02 12:58:09 +01:00
Paul Bellamy f28fe80861 Merge pull request #814 from weaveworks/revert-813-proxy-ssl 
Revert "First pass at Proxy SSL Support"
 2015-06-02 12:49:08 +01:00
Paul Bellamy 91375a1095 Revert "First pass at Proxy SSL Support" 2015-06-02 12:48:38 +01:00
Paul Bellamy 3586ddcd63 Merge pull request #813 from paulbellamy/proxy-ssl 
First pass at Proxy SSL Support
 2015-06-02 12:48:31 +01:00
Matthias Radestock fca02eb489 Merge pull request #706 from awh/dns_delete_individual_records 
delete individual dns records

Closes #631. Closes #654.
 2015-06-02 12:35:50 +01:00
Adam Harrison 56160e0798 Merge pull request #728 from paulbellamy/test-runner 
allow test/run_all.sh to run specific test files
 2015-06-02 12:30:15 +01:00
Matthias Radestock 0b72338400 Merge remote-tracking branch 'upstream/master' 2015-06-02 11:58:38 +01:00
Matthias Radestock cc1a5548af reduce smoke test noise 2015-06-02 11:57:49 +01:00
Matthias Radestock a8b8a0948f cosmetic(ish): banish double bracket bashism 2015-06-02 11:56:27 +01:00
Paul Bellamy e7f73900cb allow test/run_all.sh to run specific test files 2015-06-02 11:45:43 +01:00
Paul Bellamy 927d40f47f Correctly building the client cert subjectAltNames 
Still haven't managed to try it due to:
https://github.com/docker/docker/issues/8943
 2015-06-02 11:34:54 +01:00
Matthias Radestock 294aafde14 cosmetic: s/when_has_fqdn/with_container_fqdn 2015-06-02 10:52:58 +01:00
Matthias Radestock dcedf241bb simplify dns-add/remove 2015-06-02 10:52:58 +01:00
Matthias Radestock bf8e82ba16 refactor: get rid of with_full_container_id 2015-06-02 10:52:58 +01:00
Paul Bellamy 79cc2d6d25 add note about not handling args with spaces for proxy tls files 2015-06-02 10:20:06 +01:00
Adam Harrison a4582fb529 Optimise DeleteRecords when explicit values are passed 2015-06-02 10:09:26 +01:00
Adam Harrison 30447ada52 Remove obsolete methods 2015-06-02 10:09:22 +01:00
Adam Harrison 66de361fd6 Smoketest dns-remove -h 2015-06-02 09:57:28 +01:00