googlegsa / manager.v3

Google Search Appliance Connector Manager
Apache License 2.0
10 stars 10 forks source link

For file connectors the delay after completing never gets executed #237

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. Set up the connector maanger to scan a directory with only a few files
2. Configure a retry delay in the schedule for this connector
3. Start the connector

What is the expected output? What do you see instead?
The connector should run and then after processing all files be inactive for 
the amount of time specified in the retry delay.
Instead the connector keeps on polling for changed files without an break.

What version of the product are you using? On what operating system?
2.8.4 on Linux

Please provide any additional information below.

When completing the scan of a set of directories the file connector should stop 
for the amount of time specified in the schedule as retryDelayMillis. Instead 
the file connector starts the next run immediately.

This behaviour is caused by a difference in the semantic between 
QueryTraverser.runBatch() and 
DiffingConnectorTraversalManager.newDocumentList() (Which gets used by 
startTraversal() and resumeTraversal()). The former expects the later to return 
null, when there are no more unprocessed documents, where as the later never 
will return null.

The developers of the file connector seem to be aware of this issue as they 
have their own delay build into the configuration 
(getDelayBetweenTwoScansInMillis), which may be used as a workaround.

Original issue reported on code.google.com by fredkie...@gmx.de on 31 Jul 2012 at 11:08

GoogleCodeExporter commented 9 years ago
This is a known problem with the Diffing engine.  In the 2.8 versions of the 
file system connector, you could get around it by setting the 
"delayBetweenTwoScansInMillis" advanced configuration property in 
connectorInstance.xml.

File System Connector v3.0 switches from the Diffing engine to a 
Lister/Retriever model and consequently the connector pays attention to the 
Schedule, including the Retry Delay.

Original comment by Brett.Mi...@gmail.com on 31 Jul 2012 at 8:44

GoogleCodeExporter commented 9 years ago
Fixed in File System Connector v 3.0.0

Original comment by Brett.Mi...@gmail.com on 1 Oct 2012 at 5:39