Attempts to stream messages from tcp://hub.fedoraproject.org:9940 using fedmsg-tail or fedmsg-trigger is not reliable. Either messages are not received upon connecting or sporadically received.
Datagrepper is used as a comparison in both cases.
See https://apps.fedoraproject.org/datagrepper/raw?rows_per_page=1&delta=127800
To reproduce:
# fedmsg-tail or # fedmsg-trigger --command date
# fedmsg-tail
# fedmsg-trigger --command date
Expected outcome:
Notes:
As soon as possible.
Metadata Update from @pingou: - Issue assigned to pingou
So with help from @kevin we did:
ansible proxies -m shell -a 'systemctl status fedmsg-gateway'
Which shows the service running on most proxy, but the last log entry was from Sep 27th for the most recent and Aug 12th for quite a few.
So we've restarted the service:
ansible proxies -m shell -a 'systemctl restart fedmsg-gateway-3'
Let's see if that helps
Unfortunately, it seems to be the same behaviour
but I noticed your restart command included fedmsg-gateway-3 ?
Metadata Update from @smooge: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: medium-gain, medium-trouble, ops
fedmsg-tail is running fine for me here.
Can you run fedmsg-config and check the content of the endpoints dictionary?
fedmsg-config
endpoints
That's because it's the python3 version of the service ;-)
"endpoints": { "fedora-infrastructure": [ "tcp://hub.fedoraproject.org:9940" ], "relay_outbound": [ "tcp://127.0.0.1:4001" ] },
Another reproducer can be:
start a fedmsg-tail in each session but with a 10s delay between.
open another terminal session with:
fedora-messaging --conf /etc/fedora-messaging/fedora.toml consume
compare all 3.
ok. Try now? I think it's fixed?
My theory/what I fixed:
busgateway01 had a vpn ip in the old transition ip space that we used when moving from phx2 to iad2. This meant that none of the proxies could talk to it over the vpn. The only ones that worked were the ones in ia2 datacenter (since they could just talk to it directly). I fixed the IP on busgateway and restarted things and confirmed here that fedpmsg-tail seems much better.
Yep! all fixed. All 10 of my fedmsg-tail sessions are pulling in messages. Thanks!
Awesome. Sorry this happened... ;(
Metadata Update from @kevin: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.