Web site still showing most recently added packages from 4 days ago.

[chipjarred]

It's not a big deal, at least for me, but since I added my package yesterday and it's been merged, I looked at SwiftPackageIndex.com to see if it showed up as recently added. The Recently Added still shows packages from 4-7 days ago. The Recent Releases does seem to up to date. Since there are no open issues, I thought I'd mention it just in case no one else had noticed. I could see it being more of an issue for developers actively wanting to get as much exposure for their package as they can.

[finestructure]

That's odd and it seems to coincide with our move of the backend to new hosting. It's not clear to me how that would affect this but there certainly seems to be something wrong. Thanks for reporting it!

[finestructure]

Just for reference, this is the current view:

[finestructure]

Our RSS feed shows the same:

[finestructure]

Staging picked this up correctly:

[finestructure]

Reconcile is failing to connect to the db:

[ INFO ] Reconciling ... [component: reconcile]
[ ERROR ] Channel error caught. [psql_connection_id: EBB34917-47DC-40F9-8EF4-207ECC928E9B, psql_error: uncleanShutdown]
[ ERROR ] Channel error caught. Closing connection. [psql_connection_id: EBB34917-47DC-40F9-8EF4-207ECC928E9B, psql_error: PSQLError(base: PostgresNIO.PSQLError.Base.connectionError(underlying: NIOSSL.NIOSSLError.uncleanShutdown))]

[finestructure]

This is the case for all services running on p3. It doesn't seem to be able to connect to the db.

Edit: except for app_server, which does seem to be able to connect.

[finestructure]

What's odd is that app_server is connecting just fine from the job on p3:

[ INFO ] GET /marinofelipe/swift-package-info [component: server, request-id: 858996AA-35CF-4962-9648-31D1492F13E1]
[ INFO ] POST /api/versions/78A72B11-A475-478B-9189-F3AD5B3411DC/builds [component: server, request-id: B6103D4D-7C59-44F4-BAFD-6F6C7FA543FA]

I've restarted p3 but the error persists and I've now restarted it again in order to drain the processing jobs off it. That should restore functionality but it's not a solution, because any deploy or restart of another node will bring one or more jobs back to it.

p3 is the node I had to rebuild after disk resizing failed, so it might be related to the manual process that entailed. However, that should also affect connectivity from app_server.

[finestructure]

The new packages are now showing up:

I'll resolve this issue and track the p3 issue separately. Thanks again for the bug report, @chipjarred !