Hi Jeff - Your questions are very Cloudera specific. It is inappropriate for us to discuss these things in the Apache community.
I can tell you that there are multiple reasons that admins and distros provide multiple topologies in Apache Knox. One of the most common is to provide access to the same cluster and to the same services sometimes via different authentication mechanisms. The cdp-proxy-api topology is the one to use with HTTP Basic authentication. One of things that makes this a little confusing is that the UIs often require their API's to be added to the topology as well so that they can make calls to their backend. The same backend is sometimes made available in a topology intended for only APIs and the UI counterpart will be absent. Any more information for CDP needs to be acquired through company support, documentation, etc. thanks, --larry On Thu, Sep 9, 2021 at 8:49 PM Jeffrey Rodriguez <[email protected]> wrote: > Though we support Spark History server through Knox Spark History UI > service. I can reach out through a browser and do SSO authentication and > access /v1/api/applications and it will return a json body. This REST Api > as documented in > https://spark.apache.org/docs/latest/monitoring.html > The issue is that I need to use curl or a cli client , when I try the hs > url with the rest endpoint I got an authentication error since the topology > for UIs uses SSO and I am trying basic authentication instead. > Any suggestions? Would I have to add to my cdp-proxy-api? CDE have a jwt > token option where I can get a bear token but i noticed cdp doesn’t provide > it. > I am reaching support for public cdp too. I would hate to make updates on > our cloud conf for Knox so any help would be appreciated. > On a different question do you know if Knox or webhdfs support abfs:// > files? > Thanks again I have already most of the info I need thanks to Knox user > group. > Jeff Rodriguez >
