Abhilash Perla created KNOX-2709:
------------------------------------
Summary: Retry is not happening for webhdfs api in Knox HA provider
Key: KNOX-2709
URL: https://issues.apache.org/jira/browse/KNOX-2709
Project: Apache Knox
Issue Type: Bug
Reporter: Abhilash Perla
Attachments: cdp-proxy-api-1.xml, cdp-proxy-api.xml, gateway-1.log,
gateway-audit-1.log, gateway-audit.log, gateway.log,
image-2022-02-24-19-59-38-998.png, image-2022-02-24-20-00-10-074.png,
image-2022-02-24-20-01-26-352.png
When validating retry parameters (maxRetryAttempts , retrySleep) in Knox HA
provider observed that no retries are taking place for webhdfs when Namenode is
in safemode while in documentation it says the retry should take place for this
case ( Doc link :
[https://knox.apache.org/books/knox-1-1-0/user-guide.html#WebHDFS+HA] )
Observed only failover is happening in this case
*Steps taken:*
1. Turned on safe mode for Hdfs
2. Did a curl put request through webhdfs api.
*Executed command:*
curl -i -k -u knoxui:knoxui -X PUT
'[https://quasar-zyumtb-1.quasar-zyumtb.root.hwx.site:8443/gateway/cdp-proxy-api/webhdfs/v1/user/guest/example/README1?op=CREATE'|https://quasar-zyumtb-1.quasar-zyumtb.root.hwx.site:8443/gateway/cdp-proxy-api/webhdfs/v1/user/guest/example/README1?op=CREATE%27]
The output in Knox/gateway/gateway.log:
!image-2022-02-24-19-59-38-998.png!
The output in gateway-audit.log:
!image-2022-02-24-20-00-10-074.png!
The topology configurations(Snippet from cdp-proxy-api.xml):
!image-2022-02-24-20-01-26-352.png|width=603,height=42!
*Expected Output:*
Retry should happen 300 times with interval between each retry should be 1second
*Actual Output:*
In the above case failover attempts is set to 6 .
So 6 times failover happened but retry is not happening.
Note:
Added the Topology and log files of both instances in attatchments
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
