and not all the help, version, etc, etc docker runs.
In particular this allows passing `--restart=...` in
WEAVESCOPE_DOCKER_ARGS. Previously that would break since the
non-probe/app docker runs supply `--rm` (as they should) which is
incompatible with restart settings.
74dc626 Merge pull request #108 from weaveworks/disable-apt-daily
b4f1d91 Merge pull request #107 from weaveworks/docker-17-update
7436aa1 Override apt daily job to not run immediately on boot
7980f15 Merge pull request #106 from weaveworks/document-docker-install-role
f741e53 Bump to Docker 17.06 from CE repo
61796a1 Update Docker CE Debian repo details
0d86f5e Allow for Docker package to be named docker-ce
065c68d Document selection of Docker installation role.
3809053 Just --porcelain; it defaults to v1
11400ea Merge pull request #105 from weaveworks/remove-weaveplugin-remnants
b8b4d64 remove weaveplugin remnants
35099c9 Merge pull request #104 from weaveworks/pull-docker-py
cdd48fc Pull docker-py to speed tests/builds up.
e1c6c24 Merge pull request #103 from weaveworks/test-build-tags
d5d71e0 Add -tags option so callers can pass in build tags
8949b2b Merge pull request #98 from weaveworks/git-status-tag
ac30687 Merge pull request #100 from weaveworks/python_linting
4b125b5 Pin yapf & flake8 versions
7efb485 Lint python linting function
444755b Swap diff direction to reflect changes required
c5b2434 Install flake8 & yapf
5600eac Lint python in build-tools repo
0b02ca9 Add python linting
c011c0d Merge pull request #79 from kinvolk/schu/python-shebang
6577d07 Merge pull request #99 from weaveworks/shfmt-version
00ce0dc Use git status instead of diff to add 'WIP' tag
411fd13 Use shfmt v1.3.0 instead of latest from master.
31d069d Change Python shebang to `#!/usr/bin/env python`
git-subtree-dir: tools
git-subtree-split: 74dc626b6de3ffb38591510f7cb7bc2db33743c4
This guarantees that the output won't contain bare nodes containing
just an id and topology, as produced by ipToNode. Previously this was
ensured by calling convention, now it's ensured by construction.
When the scope-app restarts, it no longer has a
reference to the previous node set. Therefore,
the delta update adds *all* nodes but does not
remove legacy ones.
`reset==true` tells the frontend to start fresh.
Fixes#2708
If the scope-app API unexpectedly restarts, it has no report
at hand (until it gets one from the probe) and sends node
count 0 to the frontend for all topologies. Once the report
arrives, it will send the proper count.
What happened was the frontend did hide Processes for a short
time till the node count recovered. This moved the topology
selection to the always visible Containers (hide_if_empty == false)
while keeping the graph as is.
Once the node count recovers, Processes comes back but the
selection is still at Containers.
We now keep the selected topology visible at all time even if
the API returns a node count of 0. This recovers nicely when
the correct node counts come in. Once the user selects a different
topology while and a backend response arrives, it disappears.
Fixes#2646
...when initialising eBPF-based connection tracking.
Previously we were ignoring all eBPF events until we had gathered the
existing connections. That means we could a) miss connections created
during the gathering, and b) fail to forget connections that got
closed during the gathering.
The fix comprises the following changes:
1. pay attention to eBPF events immediately. That way we do not
miss anything.
2. remember connections for which we received a Close event during the
initalisation phase, and subsequently drop gathered existing
connections that match these. That way we do not erroneously consider
a gathered connection as open when it got closed since the gathering.
3. drop gathered existing connections which match connections detected
through eBPF events. The latter typically have more / current
metadata. In particular, PIDs can be missing from the former.
Fixes#2689.
Fixes#2700.
Without synchronisation, the isDead() call might return a stale value,
delaying deadness detection potentially indefinitely.
Without the guards / idempotence in .stop(), invoking stop() more than
once could cause a panic, since tracer.Stop() closes a channel (which
panics on a closed channel). Multiple stop() invocations are rare, but
not impossible.
when we got an fd install event but the pid was dead by time we
processed it, we would fail to remove the watcher for that pid from
the fdinstall_pids table.
This is a minor, and bounded, leak, since the table only contains pids
that were alive when we initialized ebpf. And this change only plugs
that leak very partially, since we will never remove pids that die
while sitting in accept().
The package version is irrelevant for the build process
and is not read anywhere.
The package is not published and causes confusion if the
bump is forgotten.
We defer starting the ebpf tracer until we've set the global var which
is referenced by the callback functions. Previously the var could be
unset when the callbacks are invoked, resulting in a segfault.
Fixes#2687.
