Parametrize and increase default of swift-proxy node_timeout
authorJiri Stransky <jistr@redhat.com>
Tue, 21 Jun 2016 09:37:28 +0000 (11:37 +0200)
committerJiri Stransky <jistr@redhat.com>
Tue, 21 Jun 2016 09:37:28 +0000 (11:37 +0200)
We've had problems uploading images on slow virtualized environments,
where the requests to swift would time out. The issue has been seen both
on undercloud and overcloud. The default timeout is 10 seconds, the
issue has been reportedly reproduced on undercloud with node_timeout as
high as 30 seconds, but not with 60 seconds yet. Set the default timeout
to 60 seconds on overcloud too.

Change-Id: I7d486cf4dc9768ddbf71ab71e92db8d2ef29978e
Closes-Bug: #1594725

puppet/services/swift-proxy.yaml

index a86aeaf..930b9e3 100644 (file)
@@ -17,6 +17,10 @@ parameters:
     description: The password for the swift service account, used by the swift proxy services.
     type: string
     hidden: true
+  SwiftProxyNodeTimeout:
+    default: 60
+    description: Timeout for requests going from swift-proxy to swift a/c/o services.
+    type: number
   SwiftWorkers:
     default: 0
     description: Number of workers for Swift service.
@@ -36,6 +40,7 @@ outputs:
         swift::proxy::authtoken::auth_uri: {get_param: [EndpointMap, KeystoneInternal, uri]}
         swift::proxy::authtoken::identity_uri: {get_param: [EndpointMap, KeystoneAdmin, uri_no_suffix]}
         swift::proxy::authtoken::admin_password: {get_param: SwiftPassword}
+        swift::proxy::node_timeout: {get_param: SwiftProxyNodeTimeout}
         swift::proxy::workers: {get_param: SwiftWorkers}
         swift::keystone::auth::public_url: {get_param: [EndpointMap, SwiftPublic, uri]}
         swift::keystone::auth::internal_url: {get_param: [EndpointMap, SwiftInternal, uri]}