21 Jan 2023 |
hexa | nah, looks like that completed | 21:47:17 |
22 Jan 2023 |
K900 | It got oomkilled I think | 08:00:24 |
K900 | So I just restarted it | 08:00:29 |
m1cr0man | In reply to @hexa:lossy.network
(finished: must succeed: curl --data '{"host": "acme.test", "addresses": ["192.168.1.1"]}' http://192.168.1.3:8055/add-a, in 0.24 seconds) client # curl: (7) Failed to connect to acme.test port 15000 after 88 ms: Couldn't connect to server client # curl: (7) Failed to connect to acme.test port 15000 after 88 ms: Couldn't connect to server
Yeah that looks fine | 18:53:19 |
m1cr0man | I suppose OOMkill could be the culprit actually.. this test starts like 4 vms iirc. Client/dnsserver/webserver/acme server. I don't imagine many other tests have as many VMs | 18:54:50 |
hexa | maybe allocate more memory for the test | 19:34:30 |
hexa | * maybe allocate more memory for the test then | 19:34:36 |
m1cr0man | is that possible? | 22:06:26 |
31 Jan 2023 |
Winter (she/her) | In reply to @m1cr0man:m1cr0man.com is that possible? virtualisation.memorySize , bytes. (default is 1024.) | 00:53:58 |
m1cr0man | Does that increase the ram for each node or for the encapsulating VM running the suite? | 01:03:30 |
Winter (she/her) | In reply to @m1cr0man:m1cr0man.com Does that increase the ram for each node or for the encapsulating VM running the suite? There's no encapsulating VM. Each node is run as its own VM. | 01:07:43 |
m1cr0man | Right I see, see I think the issue is that whatever the test suite is running on is running out of ram. | 01:08:21 |
Winter (she/her) | let me poke the operator of that specific machine | 01:09:01 |
m1cr0man | I already did that 103-run test a while ago and it was grand so I don't think the nodes are running out | 01:09:12 |
m1cr0man | Alright thanks 🙂 lmk if there's something obvious | 01:09:37 |
Winter (she/her) | poked them in #infra:nixos.org | 01:09:57 |
m1cr0man | Actually is there system performance dashboards we can correlate against test failure? | 01:09:59 |
Winter (she/her) | https://monitoring.nixos.org/grafana/ might have something | 01:10:41 |
m1cr0man | Cool I'll check that out tomorrow | 01:11:05 |
2 Feb 2023 |
Winter (she/her) | m1cr0man: Would you say the best way to guide users wrt DynamicUser services and permissions would be to have them set SupplementalGroups to whatever owns the given cert? | 15:18:49 |
m1cr0man | Yep | 15:19:25 |
hexa | https://hydra.nixos.org/build/207980199 acme 😄 | 17:44:04 |
hexa | https://hydra.nixos.org/log/lbyjk7n05hk7s9mhccrh4h1jzs470lkl-vm-test-run-acme.drv | 17:44:29 |
hexa | restarting | 17:44:32 |
K900 | Saved the log to https://termbin.com/nrjp | 17:45:03 |
hexa | thanks | 17:45:23 |
hexa | probably as helpful as ever | 17:45:32 |
raitobezarius | In reply to @winterqt:nixos.dev m1cr0man: Would you say the best way to guide users wrt DynamicUser services and permissions would be to have them set SupplementalGroups to whatever owns the given cert? I personally do that | 17:58:55 |
Winter (she/her) | In reply to@hexa:lossy.network probably as helpful as ever you'd be right :) ``` | 22:42:18 |
Winter (she/her) | In reply to@hexa:lossy.network probably as helpful as ever * you'd be right :) webserver: waiting for unit acme-finished-http.example.test.target
Test "Can request certificate with Lego's built in web server" failed with error: "unit "acme-finished-http.example.test.target" is inactive and there are no pending jobs" | 22:42:21 |