Open skakee opened 3 years ago
Carlos from LACNIC here. There is a very weird issue that seems to recur every 30 days or so with octo and our LACNIC repo that causes Octo's cache to go bad and stop validating the repo. If you stop octo and clear the cache it works again.
Neither routinator, fort or rpki-client report any issues. I've been running octo myself and noticed the failure but couldn't identify any significant log message.
I'd would love to find out what causes this. So any information is welcome.
not sure if anyone at Cloudflare saw the issue. I'll check but I need to setup an environment from scratch so this might take some time. Probably the thread stalls but would need to confirm with tracing.
@carlosm3011 : do you have the timings of when it happened?
The failure began on May 24th at 12.13pm UTC-3. In my measurement rig it lasted for about two hours until I wiped out the on-disk cache (accidentally actually, I destroyed the container instead of re-starting it :-) ).
On 27 May 2021, at 12:44, Louis @.***> wrote:
not sure if anyone at Cloudflare saw the issue. I'll check but I need to setup an environment from scratch so this might take some time. Probably the thread stalls but would need to confirm with tracing.
@carlosm3011 https://github.com/carlosm3011 : do you have the timings of when it happened?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/cloudflare/cfrpki/issues/86#issuecomment-849740348, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAA4ARSUDKZ3YFEHR6FWTK3TPZSEDANCNFSM43OB5OFQ.
output.json generated by OctoRPKI is missing LACNIC. I see reports that other see the same issue (please see link below). https://rpki.cloudflare.com/?ohlcTa=LACNIC shows a recent issue. Can you shed some light on it?
https://www.mail-archive.com/nanog@nanog.org/msg112880.html