Closed jgallagher closed 1 year ago
@jgallagher I believe the ddmd
build from #58 should fix this.
@rcgoodfellow Using the ddmd
build from #58 with the latest omicron main dendrite commit (963bcd784e4385cbb45ecfba548ac4b1abb6edc1
, which per @internet-diglett matches 8067c39088482ada72b005cb394270815d181bee
), I'm seeing a different error in the mg-ddm
logs:
Dec 28 00:47:16.146 INFO [53] nbr is fe80::aa40:25ff:fe04:157@unknown server
Dec 28 00:47:16.147 INFO [53] exchange: listening on [fe80::aa40:25ff:feec:f2c6]:56797
Dec 28 00:47:16.147 INFO waiting for exchange server to start
Dec 28 00:47:18.149 WARN [53] exchange pull: hyper error: error trying to connect: tcp connect error: Connection refused (os error 146)
Dec 28 00:47:18.150 INFO sending 1 routes to dendrite
Dec 28 00:47:18.157 ERRO [53] add system route: doesn't match pattern "(^[qQ][sS][fF][pP](([0-9])|([1-2][0-9])|(3[0-1]))$)|(^[rR][eE][aA][rR](([0-9])|([1-2][0-9])|(3[0-1]))$)|(^[iI][nN
][tT]0$)"
Dec 28 00:47:18.157 INFO removing routes 0 from dendrite
Ah i see, sorry, one moment ....
@jgallagher I believe I've fixed this - if your lap times are long in trying these changes out, I can spin up my ddmd/dendrite development environment and run this through tests.
Within the switch zone on
madrid
, we observed that even though magehemite had found prefixes for peers (unknown
here is actually the hostname of the peer):there were no routes set up:
@rcgoodfellow noted this in the
mg-ddm
logs that indicates the culprit: