This runner is in my homelab while we muse about a better, more
permanent home for these tests.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
This makes integration tests pull pristine VM images from Amazon S3 if
they don't exist on disk. If the S3 fetch fails, it will fall back to
grabbing the image from the public internet. The VM images on the public
internet are known to be updated without warning and thusly change their
SHA256 checksum. This is not ideal for a test that we want to be able to
fire and forget, then run reliably for a very long time.
This requires an AWS profile to be configured at the default path. The
S3 bucket is rigged so that the requester pays. The VM images are
currently about 6.9 gigabytes. Please keep this in mind when running
these tests on your machine.
Documentation was added to the integration test folder to aid others in
running these tests on their machine.
Some wording in the logs of the tests was altered.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
The image downloads can take a significant amount of time for the tests.
This creates a new test that will download every distro image into the
local cache in parallel, optionally matching the distribution regex.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
I've run into a couple issues where the tests time out while a VM image
is being downloaded, making the cache poisoned for the next run. This
moves the hash checking into its own function and calls it much sooner
in the testing chain. If the hash check fails, the OS is redownloaded.
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Most of the time qemu will output nothing when it is running. This is
expected behavior. However when qemu is unable to start due to some
problem, it prints that to either stdout or stderr. Previously this
output wasn't being captured. This patch captures that output to aid in
debugging qemu issues.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Previously this built the binaries for every distro. This is a bit
overkill given we are using static binaries. This patch makes us only
build once.
There was also a weird issue with how processes were being managed.
Previously we just killed qemu with Process.Kill(), however that was
leaving behind zombies. This has been mended to not only kill qemu but
also waitpid() the process so it doesn't become a zombie.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
The OpenSUSE 15.1 image we are using (and conseqentially the only one
that is really available easily given it is EOL) has cloud-init
hardcoded to use the OpenStack metadata thingy. Other OpenSUSE Leap
images function fine with the NoCloud backend, but this one seems to
just not work with it. No bother, we can just pretend to be OpenStack.
Thanks to Okami for giving me an example OpenStack configuration seed
image.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Arch is a bit of a weirder distro, however as a side effect it is much
more of a systemd purist experience. Adding it to our test suite will
make sure that we are working in the systemd happy path.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
This distro is about to be released. OpenSUSE has historically had the
least coverage for functional testing, so this may prove useful in the
future.
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Instead of testing all the VMs at once when they are all ready, this
patch changes the testing logic so that the vms are tested as soon as
they register with testcontrol. Also limit the amount of VM ram used at
once with the `-ram-limit` flag. That uses a semaphore to guard resource
use.
Also document CentOS' sins.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
If you set `-distro-regex` to match a subset of distros, only those
distros will be tested. Ex:
$ go test -run-vm-tests -distro-regex='opensuse'
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Don't try to do heuristics on the name. Use the net/interfaces package
which we already have to do this sort of stuff.
Fixes#2011
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Instead of pulling packages from pkgs.tailscale.com, we should use the
tailscale binaries that are local to this git commit. This exposes a bit
of the integration testing stack in order to copy the binaries
correctly.
This commit also bumps our version of github.com/pkg/sftp to the latest
commit.
If you run into trouble with yaml, be sure to check out the
commented-out alpine linux image complete with instructions on how to
use it.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Previously we spewed a lot of output to stdout and stderr, even when
`-v` wasn't set. This is sub-optimal for various reasons. This patch
shunts that output to test logs so it only shows up when `-v` is set.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
Instead of relying on a libvirtd bridge address that you probably won't
have on your system.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
This will spin up a few vms and then try and make them connect to a
testcontrol server.
Updates #1988
Signed-off-by: Christine Dodrill <xe@tailscale.com>
* Added new Addresses / AllowedIPs fields to testcontrol when creating new &tailcfg.Node
Signed-off-by: Simeng He <simeng@tailscale.com>
* Added single node test to check Addresses and AllowedIPs
Signed-off-by: Simeng He <simeng@tailscale.com>
Co-authored-by: Simeng He <simeng@tailscale.com>
Only minimal tailscale + tailscaled for now.
And a super minimal in-memory logcatcher.
No control ... yet.
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Importing the non-main package was missing some dependencies that
"go mod tidy" would then cleanup. Also added a non-ignore build tag to
avoid other tools getting upset about importing a main package.
Signed-off-by: Filippo Valsorda <hi@filippo.io>
Upstream wireguard-go decided to use errors.Is(err, net.ErrClosed)
instead of checking the error string.
It also provided an unsafe linknamed version of net.ErrClosed
for clients running Go 1.15. Switch to that.
This reduces the time required for the wgengine/magicsock tests
on my machine from ~35s back to the ~13s it was before
456cf8a3765948d6f1992162993eaf3844371592.
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
Use tb.Cleanup to simplify both the API and the implementation.
One behavior change: When the number of goroutines shrinks, don't log.
I've never found these logs to be useful, and they frequently add noise.
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
If any goroutine continues to use the logger in TestLocalLogLines
after the test finishes, the test panics.
The culprit for this was wireguard-go; the previous commit fixed that.
This commit adds suspenders: When the test is done, make logging calls
into no-ops.
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
Brad Fitzpatrick
Christine Dodrill
Simeng He
simenghe
Josh Bleecher Snyder
Josh Bleecher Snyder
Filippo Valsorda
David Crawshaw
Wendi Yu