fix: do not abort when unrelated connections are not yet loaded

[pdecat]

This PR resolves issues when connection configurations are created dynamically concurrently to executing queries.

Symptoms are that an unrelated connection that hasn't completed loading breaks other connections:

2024-04-05 12:05:19.955 UTC [INFO]  hub: goFdwBeginForeignScan, connection 'aws_000000000000', table 'aws_ec2_transit_gateway', explain: false
2024-04-05 12:05:19.957 UTC [WARN]  hub: pluginHub.GetIterator FAILED: failed to attach to plugin process for 'aws@0.133.0': no connection config loaded for 'aws_222222222222'
2024-04-05 12:05:19.957 UTC [299] ERROR:  failed to attach to plugin process for 'aws@0.133.0': no connection config loaded for 'aws_222222222222'
2024-04-05 12:05:19.957 UTC [299] STATEMENT:  SELECT 'APIGATEWAY' AS subtype, row_to_json(r) AS data FROM (SELECT
[...SNIP...]
2024-04-05 12:05:19.958 UTC [INFO]  hub: goFdwAbortCallback
2024-04-05 12:05:19.958 UTC [INFO]  hub: Hub Abort
2024-04-05 12:05:20.023 UTC [INFO]  hub: goFdwIterateForeignScan returned empty row - this scan complete (0xc00226f080)
2024-04-05 12:05:20.025 UTC [INFO]  hub: goFdwIterateForeignScan calling pluginHub.StartScan, table 'aws_account' (0xc00226f680)
2024-04-05 12:05:20.025 UTC [INFO]  hub: StartScan for table: aws_account, cache enabled: false, iterator 0xc00226f680, 0 quals (1712318705836)
2024-04-05 12:05:20.026 UTC [INFO]  hub: goFdwIterateForeignScan returned empty row - this scan complete (0xc00226f680)
2024-04-05 12:05:20.028 UTC [INFO]  hub: goFdwReScanForeignScan, connection '', table 'aws_ec2_transit_gateway'
2024-04-05 12:05:20.028 UTC [INFO]  hub: goFdwBeginForeignScan, connection 'aws_111111111111', table 'aws_ec2_transit_gateway', explain: false
2024-04-05 12:05:20.030 UTC [WARN]  hub: pluginHub.GetIterator FAILED: failed to attach to plugin process for 'aws@0.133.0': no connection config loaded for 'aws_222222222222'
2024-04-05 12:05:20.030 UTC [INFO]  hub: .
******************************************************

                steampipe postgres fdw shutdown

******************************************************

This problem seems to happen more frequently when connections are slower to load, most notably when limiters are configured.

See https://turbot-community.slack.com/archives/C01UECB59A7/p1712324097931309

For this change to be effective, the FDW has to be rebuilt with:

go mod edit -replace="github.com/turbot/steampipe=github.com/pdecat/steampipe @0a5b8a0928292f08ad842a63211188306e9998de"

[pdecat]

Submitted a reproduction case: https://github.com/pdecat/steampipe-4233-repro

[kaidaguerre]

hi @pdecat
sorry for the long delay on this one. I actually spent a bunch of time digging into this one to try and identify the underlying reasons, but I failed/ran out of time and had to move on to other stuff.

Came back to this today and I think it's a fine tactical fix for now. I'd still like to get to the bottom of it but I'll merge in the meantime.

[pdecat]

Hi @kaidaguerre, no problem for the delay, been using my own fork until now :)

Thanks! 🎉