[UNDERTOW-1759][UNDERTOW-1523] Initial GitHub actions test matrix

[luck3y]

This builds on top of Stuart's initial version, but splits the tests to use a matrix and test each module in parallel.

Should also address: #901

Jiras: https://issues.redhat.com/browse/UNDERTOW-1759 and https://issues.redhat.com/browse/UNDERTOW-1523

[luck3y]

FYI, there are likely too many combinations here, but having some extra ones is useful for looking at some of the failures, for example the JDK 11 Linux ipv6 tests pass, while the equivalent ipv4 ones do not. Once the suite is stabilized, I think we could eliminate some combinations.

[undertow-pull-request]

Hello, luck3y. I'm waiting for one of the admins to verify this patch with /ok-to-test in a comment.

[luck3y]

There are a couple of notable things, including the GHA hosts files being filled with malicious domains mapped back to localhost, which causes some of the failures. (The test setup overwites /etc/hosts to address this.)

There are still common failures that do appear to consistently occur (I think there are tickets open for most of those already.)

See here for example: https://github.com/luck3y/undertow/pull/2/checks

[luck3y]

One other observation. These hosts all have:

`docker0: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500
inet 172.17.0.1 netmask 255.255.0.0 broadcast 172.17.255.255
ether 02:42:3f:4e:17:df txqueuelen 0 (Ethernet)
RX packets 0 bytes 0 (0.0 B)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 0 bytes 0 (0.0 B)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 10.1.0.4 netmask 255.255.0.0 broadcast 10.1.255.255
inet6 fe80::20d:3aff:fe9a:96d3 prefixlen 64 scopeid 0x20
ether 00:0d:3a:9a:96:d3 txqueuelen 1000 (Ethernet)
RX packets 158412 bytes 224141913 (224.1 MB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 7482 bytes 876141 (876.1 KB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536
inet 127.0.0.1 netmask 255.0.0.0
inet6 ::1 prefixlen 128 scopeid 0x10
loop txqueuelen 1000 (Local Loopback)
RX packets 493 bytes 52460 (52.4 KB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 493 bytes 52460 (52.4 KB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0`

but lookups for the hostname show two answers:

`; <<>> DiG 9.11.3-1ubuntu1.12-Ubuntu <<>> fv-az55
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 17373
;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 65494
;; QUESTION SECTION:
;fv-az55. IN A

;; ANSWER SECTION:
fv-az55. 0 IN A 10.1.0.4
fv-az55. 0 IN A 172.17.0.1
`
I'm not sure if a hostname other than localhost is used for anything, but the eth0 one looks likely to be the one that should be used in the case it was looked up. This could also explain why tests failing on ipv4 then pass on ipv6.

[stuartwdouglas]

1609b26 should fix the HTTP continue test failure

[stuartwdouglas]

@fl4via I think we should merge this, it will make it a lot easier to stabilise the test suite.

[luck3y]

FYI, if / when this gets merged, I"ll pause the pull-player.

[stuartwdouglas]

I have done some more work on this today, both in fixing flakey tests and I also expanded the matrix somewhat: https://github.com/stuartwdouglas/undertow/actions/runs/186634583

Instead of -Pproxy it now runs each protocol as a seperate matrix param. I think this is overkill long term, but I think for now it will help give insight into exactly what is failing.

[fl4via]

@luck3y and @stuartwdouglas thanks, I'm reviewing everything, and working on merging it

[fl4via]

I'll be updating the commit messages to include corresponding Jiras, I'll add a message here when all this is done. I'll also include recent fixes from @stuartwdouglas

[luck3y]

Thanks @fl4via

FYI @stuartwdouglas I just noticed https://github.com/stuartwdouglas/undertow/runs/922224215?check_suite_focus=true#step:10:660 in your branch, likely we'll need to update the pam limits as well as fsmax. Maybe that will help with some of the other issues too.

[luck3y]

https://stackoverflow.com/questions/40617651/cannot-change-the-maximum-open-files-per-process-with-sysctl

…

On Wed, Jul 29, 2020, 02:40 Flavia Rainone ***@***.***> wrote: I'll be updating the commit messages to include corresponding Jiras, I'll add a message here when all this is done. I'll also include recent fixes from @stuartwdouglas <https://github.com/stuartwdouglas> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#912 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC7HACJAFWQUMFM374CPV3R57G5PANCNFSM4PJ4INQA> .

[luck3y]

Err, sorry, meant to mail that to myself :(

…

On Wed, Jul 29, 2020, 04:21 Ken Wills ***@***.***> wrote: https://stackoverflow.com/questions/40617651/cannot-change-the-maximum-open-files-per-process-with-sysctl On Wed, Jul 29, 2020, 02:40 Flavia Rainone ***@***.***> wrote: > I'll be updating the commit messages to include corresponding Jiras, I'll > add a message here when all this is done. I'll also include recent fixes > from @stuartwdouglas <https://github.com/stuartwdouglas> > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#912 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAC7HACJAFWQUMFM374CPV3R57G5PANCNFSM4PJ4INQA> > . >

[fl4via]

@luck3y this PR is ready to be merged. Since I edited it, adding Jiras, etc, could you please do a review before I merge it? Any future work can be continued in a new, separate PR.

[luck3y]

@fl4via LGTM, thanks. This will get us up and running on github, so +1 from me.

[luck3y]

FYI, I've paused the undertow TeamCity / pull-player integration now. The jobs are all still there, and they can be re-enabled at any point.