Both dbproxy1018 and dbproxy1019 are in row C and apparently in the same rack, C5. When the row is down for maintenance, the wikireplicas will be inaccessible to cloud users of any kind (VPS, Toolforge or PAWS).
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | cmooney | T291627 Packet Drops on Eqiad ASW -> CR uplinks | |||
| Resolved | cmooney | T284592 Adjust egress buffer allocations on ToR switches | |||
| Resolved | None | T286065 Switch buffer re-partition - Eqiad Row C | |||
| Resolved | aborrero | T286614 Communicate wikireplicas outage and healthcheck the system after Eqiad Row C network changes |
I'll leave this ping test running for the time of the operation:
aborrero@clouddb-wikireplicas-proxy-1:~$ ping wikireplicas-b.wikimedia.org PING wikireplicas-b.wikimedia.org (208.80.154.243) 56(84) bytes of data. 64 bytes from wikireplicas-b.wikimedia.org (208.80.154.243): icmp_seq=1 ttl=60 time=0.442 ms 64 bytes from wikireplicas-b.wikimedia.org (208.80.154.243): icmp_seq=2 ttl=60 time=0.411 ms [..] aborrero@clouddb-wikireplicas-proxy-2:~$ ping wikireplicas-a.wikimedia.org PING wikireplicas-a.wikimedia.org (208.80.154.242) 56(84) bytes of data. 64 bytes from wikireplicas-a.wikimedia.org (208.80.154.242): icmp_seq=1 ttl=60 time=0.446 ms 64 bytes from wikireplicas-a.wikimedia.org (208.80.154.242): icmp_seq=2 ttl=60 time=0.439 ms [..]
I'll report back what I see, packets loss etc.