2021-02-05 01:50:15 UTC - Dominic Kim: It's not exactly between runc and containerd but we internally did compare the performance of runc vs Docker(without runc). We got around 6~9 times more performance with runc enabled. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612489815089700?thread_ts=1612471046.089600&cid=C3TPCAQG1 ---- 2021-02-05 09:39:25 UTC - David Williams: I am assigned to use OpenWhisk to build a private cloud platform for my employee. This private cloud platform will provide almost the same features as IBM cloud platform. Can I achieve it? I would appreciate it if you could provide me with a tutorial how to do it? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612517965092800?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 14:26:51 UTC - Rodric Rabbah: Yes of course. @Michele Sciabarra FYI https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612535211093200?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 14:27:20 UTC - Rodric Rabbah: Are you deploying on bare metal of kubernetes? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612535240093700?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 14:48:32 UTC - David Williams: on a bare metal server https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612536512093900?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 14:50:08 UTC - Rodric Rabbah: the `ansible` deployment instructions in the repo are the best for that.
if this is a dev/test environment, you can even considering using the standalone openwhisk deployment, which is the easiest deployment https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612536608094100?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 14:57:54 UTC - David Williams: what is the ansible deployment instructions in the repo? I find out <https://github.com/apache/openwhisk/blob/1.0.0/tools/ubuntu-setup/README.md> https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612537074094300?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 15:50:59 UTC - Rodric Rabbah: it’s linked from there <https://github.com/apache/openwhisk/blob/1.0.0/ansible/README.md> https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612540259094500?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 20:14:46 UTC - Bilal: Quick question b/c my brain hurts. What's the entry point into Openwhisk? I have a webhook coming from grafana, but I don't see my action even activating. I don't see any errors on grafana's end. I can't recall if I should be checking the invoker or the controller? or perhaps the nginx logs? This is self hosted on kubernetes. . face_with_head_bandage : Rodric Rabbah https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556086096200?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:22:18 UTC - Bilal: I've done something similar @David Williams on a self-hosted kubernetes cluster. However I don't have many of the IBM cloud features that are nice such as the web UI. I mainly interact with openwhisk via command line. Mostly it's running scripts that myself and a few others have written so interacting with command line isn't that bad. Eventually I would like to add some web UI features such as allowing users to CRUD actions and view the output of their actions. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556538096300?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 20:25:51 UTC - Rodric Rabbah: for web ui - do you want a management dashboard for functions, or a developer console that bring the cli/terminal experience to the browser? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556751096600?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 20:26:11 UTC - Rodric Rabbah: i dont understand the question :sweat_smile: https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556771096800?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:26:40 UTC - Bilal: Grafana claims to have sent the payload (POST to my action) but my action didn't activate. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556800097400?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:26:48 UTC - Rodric Rabbah: can you share the url? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556808097600?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:26:57 UTC - Rodric Rabbah: dont need the API host just after it https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556817097800?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:27:08 UTC - Rodric Rabbah: you can check the nginx logs, did it receive the request? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612556828098000?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:30:38 UTC - Bilal: What's odd is that none of these logs are from today, this is what kube reports as logs of this container. Maybe I need to look elsewhere. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612557038098200?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:31:31 UTC - Bilal: `<http://IP/api/v1/web/guest/default/webhook>` is what I want grafana to hit, I've been doing it for almost a year now https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612557091098400?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:32:35 UTC - Bilal: Yesterday I added some more grafana alerts, changed my action script and now I'm having issues lol. So it's either coincidence or I broke something. What's odd is that it's about 50/50 chance whether grafana POSTs successfully to my openwhisk action. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612557155098600?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:39:37 UTC - Rodric Rabbah: url looks ok nginx errors are suspicious though - is the controller up/reachable? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612557577098900?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:46:45 UTC - Bilal: Both :slightly_smiling_face: Right now the barrier to entry for devs wanting a quick place to host scripts is too high (go download this cli, here's credentials, etc, etc) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558005099200?thread_ts=1612517965.092800&cid=C3TPCAQG1 ---- 2021-02-05 20:47:50 UTC - Bilal: Yes, controller is up, in the past hour openwhisk has executed over 3800 activations https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558070099400?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:54:30 UTC - Bilal: These nginx errors are from a few days ago, so that's not the current issue I don't think. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558470099600?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:55:24 UTC - Rodric Rabbah: weird https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558524099800?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:55:32 UTC - Rodric Rabbah: making my brain hurt now too https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558532100000?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:55:42 UTC - Bilal: I'm starting to suspect it's actually grafana failing to send out webhooks or silently failing https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558542100200?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:56:01 UTC - Bilal: Which is also, you guessed it, hosted on prem on kubernetes :slightly_smiling_face: https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558561100400?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 20:56:07 UTC - Rodric Rabbah: can you curl the route from the grafana pod, make sure it’s not some networking issue https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558567100600?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 21:00:19 UTC - Bilal: I appreciate your help on a Friday afternoon. I do have to get going at the moment so my debugging efforts will resume another time +1 : Rodric Rabbah https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612558819100800?thread_ts=1612556086.096200&cid=C3TPCAQG1 ---- 2021-02-05 21:06:16 UTC - Brendan Doyle: We're facing an issue with trying to use runc where containers are forcibly removed after every run. The problem is the same as described in this issue. Posting here for more visibility in case anyone else has faced it in the past and remembers how to work around it. <https://github.com/apache/openwhisk/issues/3711> https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612559176102600 ---- 2021-02-05 21:07:24 UTC - Rodric Rabbah: does the pause fail? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612559244102800 ---- 2021-02-05 21:07:48 UTC - Rodric Rabbah: (if pause fails, that would force the container’s removal, iirc) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612559268103200?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 21:09:52 UTC - Brendan Doyle: yea the pause fails that's the reason for the container removal. The status of runc's tracking of containers in the issue I think is closer to the root cause https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612559392103300?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 21:26:06 UTC - Rodric Rabbah: did you try to run runc pause yourself on the node (does it give more information)? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612560366103500?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 21:27:22 UTC - Brendan Doyle: should I try that from within the invoker container or on the server https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612560442103700?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 21:30:39 UTC - Rodric Rabbah: i’d do invoker first see if that fails and sheds any light https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612560639103900?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 22:49:24 UTC - Dave Grove: i put a comment in the issue. I think if the versions don’t match almost exactly you’ll get a grpc error. that’s my recollection from several years ago. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612565364104100?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 22:55:57 UTC - Brendan Doyle: Yea I double checked the versions are matching. I think I'm on to something. So what I mount in the invoker container. when I do `docker-runc list` it has a bunch of containers listed as stopped same as in the issue that don't seem to really match up. This is what my mount looks like `-v /run/docker/runtime-runc/moby:/run/runc` Now on my machine if I do `sudo runc --root /run/docker/runtime-runc/moby/ list` it returns a bunch of containers that are running or paused. I thought that our invoker container is running as root. If I do a `ps aux` all of the processes in the invoker container are root. But that seems to be possible it what is being mounted to the container is not the root file on the host https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612565757104300?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 23:20:45 UTC - Brendan Doyle: Actually it is the same list of containers on the host vs. in the container. But within the host every container is listed as stopped and then on the host some are listed and paused / running with none stopped https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612567245104700?thread_ts=1612559268.103200&cid=C3TPCAQG1 ---- 2021-02-05 23:52:22 UTC - Brendan Doyle: and all pid's are 0 within the container https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1612569142105000?thread_ts=1612559268.103200&cid=C3TPCAQG1 ----
