Some API takes a long time before returning 404

[Masahito Muroi]

When client calls API that are related to un-defined datasource, like /v1/data-sources/novas/tables, the API takes a long time before the API return 404 error to client in dse2. The waiting time is equal to timeout of rpc call, 1 min in default.

If possible, 404 error should be returned more quick or the timeout should be configurable.

[Eric K (ekcs)]

The RPC timeout can be changed here:
https://github.com/openstack/congress/blob/master/congress/dse2/dse_node.py#L248

What would be an appropriate value? 5s?

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/341904

[Masahito Muroi (muroi-masahito)]

For API call, 5s sounds fine.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/341904
Committed: https://git.openstack.org/cgit/openstack/congress/commit/?id=1b606a010d0b5ddb3cf77aca1e06e7a7bbc2ab3c
Submitter: Jenkins
Branch: master

commit 1b606a010d0b5ddb3cf77aca1e06e7a7bbc2ab3c
Author: Eric K <email address hidden>
Date: Thu Jul 14 18:18:55 2016 -0700

    API routing to intranode PE

    Implements the HA task to make API route PE requests to same-node PE

    Also combined with changes for fast failure on requesting
    non-existent service, because many changes shared.

    Basic idea is to first ping the service with a short timeout, then
    on success, send actual request with longer timeout.

    New DSE config parameters:
    dse_ping_timeout: timeout for pinging a service
    dse_long_timeout: timeout for potentially long requests such as row
        query on PE and action exec on DSD

    Closes-Bug: 1600171
    Closes-Bug: 1600011
    Closes-Bug: 1599991
    Partially implements blueprint: high-availability-design

    Change-Id: Ia31f70df89c8a92d441e2c197b23dc8cabc4e464

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/congress 4.0.0.0b3 development milestone.