| 24 Oct 2023 |
 philip.peter_10466 | That's one option I'm considering, it gets a bit tricky, because this is an embedded device with a read-only root and seperate user writable partition. So I would probably need to point it to a different config after the keys are fetched. | 10:32:35 |
 bazsi77 | I think if the handshake fails it will retry eventually | 10:33:18 |
| bazsi77 | Do you have or do you plan to have a disk buffer? | 10:33:57 |
| bazsi77 | Or just buffer to memory? | 10:34:11 |
| philip.peter_10466 | I'm wondering if I could use the existing buffering mechanic, and use the hook-commands to say "If there are no key files, create empty files" this will cause the sending to fail, but the buffer should take care of that. | 10:34:31 |
| philip.peter_10466 | I can use a disk buffer | 10:34:41 |
| philip.peter_10466 | We also log to 3 destinations, serial console, local file and remote. | 10:35:10 |
| philip.peter_10466 | So there is space for that | 10:35:18 |
| bazsi77 | If syslog-ng starts up, it will retry indefinitely until there's a successful handshake. The retry mechanism is controlled by the time-reopen() option, but probably a syslog-ng reload will trigger it immediately. | 10:36:27 |
| philip.peter_10466 | In an ideal situation the files should be there shortly after networking started, but it's not guaranteed and my most important concern is that the syslog-ng itself starts, and at least logs locally and to console | 10:36:32 |
| philip.peter_10466 | If there are no files it exits straight away (makes sense, the config is invalid), so the empty files would be a workaround to not have it exit | 10:37:47 |
| bazsi77 | I'd have to recreate this use case to see how syslog-ng reacts. But this should work, with a couple of ugly looking error messages. | 10:37:58 |
| philip.peter_10466 | Ideally I would want to change the config once we have the files, but I'm a bit worried about the complexity this might introduce. | 10:38:25 |
| philip.peter_10466 | Great, then I'm at least not running in the completely wrong direction, I'm going to try if the hook-commands work. Otherwise I see if I can populate the user partition with some emptty placeholder files. | 10:39:29 |
| philip.peter_10466 | Thank, you! this was very helpful! | 10:39:59 |
| bazsi77 | Let me know how this goes. Some tweaks are possible if it doesn't work | 10:49:49 |
 grg_2145 | Hi
I have a question on the syslog-ng --fd-limit flag,
the --help menu shows that it sets the The minimum required number of fds but the actual code uses it like so:
limit.rlim_cur = limit.rlim_max = process_opts.fd_limit_min; is it suppose to say minimum required or somethin like maximum allowed? (I can't see https://man7.org/linux/man-pages/man2/getrlimit.2.html having a minimum setting for opened descriptor) | 14:33:49 |
| grg_2145 changed their display name from grg_2145#0 to grg_2145. | 14:33:58 |
| 28 Oct 2023 |
 czp | I do not know, who is Mark, but I love this quote from him at (https://discuss.openobserve.ai/t/syslog-ingestion-for-cloud-and-dotnet-serilog-integration/2K1d0e): "While I am biased a bit from having developed a syslog-ng based product, nothing beats a true server (syslog-ng) for flexibility and scale. Fluentd/vector are OK for smaller setups, but once you go over a couple of terabytes a day in syslog data ingest, you’ll thank yourself for going with something more robust." | 15:25:52 |
| bazsi77 | It's probably a help screen bug. The point is that we can increase the number of permitted fds | 16:19:53 |
| bazsi77 | So we have at least the specified amount | 16:20:06 |
| bazsi77 | Mark works for axoflow 🙂 | 16:20:24 |
| czp | Ah, OK 🙂 | 16:53:47 |
| 30 Oct 2023 |
|  kamaljeetkaur_36800 joined the room. | 13:38:26 |
|  mjnovice. joined the room. | 16:10:08 |
| 31 Oct 2023 |
 alltilla1337 | It seems like the nightly APT package release is failing because of an expired token https://github.com/syslog-ng/syslog-ng/actions/runs/6699810013/job/18204916492#step:6:1 | 12:08:24 |
| alltilla1337 | authenticationerrordetail:Signature not valid in the specified time frame: Start [Fri, 22 Oct 2021 12:30:04 GMT] - Expiry [Sun, 22 Oct 2023 20:30:04 GMT] - Current [Mon, 30 Oct 2023 23:28:50 GMT] | 12:08:57 |
| alltilla1337 | hofione grg_2145 Could you check it please? Thanks! | 12:09:28 |
| 1 Nov 2023 |
 hofione | I guess this should be forwarded to Folti, I'll ping him | 18:59:52 |
| 2 Nov 2023 |
|  pepov | 15:00:00 |
