File Browser on Cloudera Manager is not up to date

[#1]

After working around my previous problem here, I found out that the File Browser on Cloudera Manager is not up to date, all the files & folders are old, which makes reports to not show anything useful. On the top right corner on HDFS -> File Browser -> Directory Usage page it says "The file system image was last indexed on January 26, 2017 1:55 PM". 

This is preventing me from generating disk usage reports and also analyzing statistics on Cloudera Navigator.

I' ve found this very informing article, but couldn't really helped me.

If I'm not wrong there is something wrong with namenodes are not checkpointing or not using the image they extract. Can anyone give more information about this?

As it shows in the outputs I posted below, it does checkpoint every 1 hour as in the config but how can I make Cloudera Manager, Navigator, File Browser etc. to see it?

I have 1 active 1 stand-by namenodes.

Checkpoint directory setting: 

Though, I have no "/ds1/dfs/snn" folder on stand-by namenode machine.

When I search for fsimage* on active namenode:

2015-08-19+17:54: 7878 ./ds1/dfs/snn/current/fsimage_0000000000000000279
2015-08-19+17:54  62 ./ds1/dfs/snn/current/fsimage_0000000000000000279.md5
2015-08-19+17:54  10272 ./ds1/dfs/snn/current/fsimage_0000000000000000510
2015-08-19+17:54  62 ./ds1/dfs/snn/current/fsimage_0000000000000000510.md5
2017-03-29+13:57  281926964 ./ds1/dfs/nn/current/fsimage_0000000000287939457 (today)
2017-03-29+13:57  281926964 ./ds2/dfs/nn/current/fsimage_0000000000287939457
2017-03-29+13:58  62 ./ds1/dfs/nn/current/fsimage_0000000000287939457.md5
2017-03-29+13:58  62 ./ds2/dfs/nn/current/fsimage_0000000000287939457.md5
2017-03-29+14:58  282426827 ./ds1/dfs/nn/current/fsimage_0000000000287980225
2017-03-29+14:58  282426827 ./ds2/dfs/nn/current/fsimage_0000000000287980225
2017-03-29+14:58  62 ./ds1/dfs/nn/current/fsimage_0000000000287980225.md5
2017-03-29+14:58  62 ./ds2/dfs/nn/current/fsimage_0000000000287980225.md5

on stand-by namenode:

2017-03-29+13:57  281926964 ./ds1/dfs/nn/current/fsimage_0000000000287939457
2017-03-29+13:57  62 ./ds1/dfs/nn/current/fsimage_0000000000287939457.md5
2017-03-29+13:57  281926964 ./ds2/dfs/nn/current/fsimage_0000000000287939457
2017-03-29+13:57  62 ./ds2/dfs/nn/current/fsimage_0000000000287939457.md5
2017-03-29+14:58  282426827 ./ds1/dfs/nn/current/fsimage_0000000000287980225
2017-03-29+14:58  282426827 ./ds2/dfs/nn/current/fsimage_0000000000287980225
2017-03-29+14:58  62 ./ds1/dfs/nn/current/fsimage_0000000000287980225.md5
2017-03-29+14:58  62 ./ds2/dfs/nn/current/fsimage_0000000000287980225.md5

[#2]

I would defer to someone from Cloudera on this, but i believe that Cloudera Manager saves a snapshot of your FSImage periodically as well. It is possible that it is using that snapshot for the file browser and not actually making the requests to HDFS.

[#3]

@burakongun,

I am not sure if you opened a new thread, but as suggested, the Reports Manager will download the fsimage every hour and attempt to index it.

To start, check the Reports Manager log (/var/log/cloudera-scm-headlamp) for errors getting the fsimage or indexing.

If you can't solve, open a new thread in the Cloudera Manager forum

-Ben

[#4]

@bgooley

Thanks for the tip. I checked Reports Manager and Navigator Metadata Server logs and got some insights about the problem. Appearently there's an issue with Kerberos authentication. I'll post Cloudera Manager forum if I can't solve it.