Rafi KC

@ffilz When I test a simple file create, read and delete in a loop for the same number of files with same filename, the memory keep constants. But if I...

@ffilz I reproduced again, but this time with proper symbols and a more detailed massif output. I even used an output visualizer to read. These tools suggested some memory consumption....

I ran the work load over the weekend where the memory footprint increased to 57GB [massif-test-second-run.log](https://github.com/nfs-ganesha/nfs-ganesha/files/7193832/massif-test-second-run.log). As you can see on the massif output the same backtrace consuming 52.6GB memory...

@Mr-Chengcheng Thanks for your suggestion, I'm testing it. I will report my status. Regarding the tools and command, Command used `valgrind --tool=massif --pages-as-heap=yes --time-unit=B --smc-check=all --trace-children=yes --massif-out-file=/var/log/massif-test1.out --max-stackframe=1201018400000 /usr/bin/ganesha.nfsd -F`...

@Mr-Chengcheng I repeated my testing with the open-behind xlater turned off. It didn't help. I ran the workload over the weekend and the memory reached 87GB. Here is the massif...

We are using shared storage hosted by a gluster.

@ffilz We use Gluster FSAL. And for the HA we use ctdb along with the shared storage backed by gluster(an exclusive shared storage) volume mounted via fuse.

@ffilz We don't have a VIP for each node, We only have one VIP across the cluster. That will be moved when one node is unavailable. Do you think this...

@prashanthpai , As far as I understand Metadata can only use as a tag. It cannot have different states, or in other words, it can not perform collective actions of...

The discussion were moved to the issue https://github.com/gluster/glusterd2/issues/1094. Initially the PR was submitted as labels being a general use case api where it can be used for volumes,snapshots, etc. But...