I also run ganesha active/active with ctdb, my versions are quite outdated now, haven't done much tuning but I'll share some thoughts

I'm assuming you're using the ceph FSAL with Ganesha rather than exporting a cephfs kernel mount.

Increasing the ceph client cache size made the biggest difference for me, that's the  client_oc_size parameter, you set that in the [client] section of your ceph.conf on the Ganesha servers. I have all the Ganesha caching turned off in my Ganesha.conf (don't have access currently to post parameters)

If you have any reasonably fast local storage on one of the Ganesha servers, I would export a dir using the VFS FSAL and mount that on your user's terminal and test their workload. If there's no improvement, that would suggest the bottleneck is the nfs implementation rather than ceph.

Or try and reproduce the workload on a ceph kernel mount, if performance is still poor that would suggest the issue is with your cluster.

In terms of benchmarking I typically use good old fio

Let me know how you get on





On Fri, 1 Oct 2021, 11:09 , <zlaymonk@gmail.com> wrote:
Hi,

We have a 2-node NFS ganesha cluster (using CTDB), that I recently inherited administration for.

Most of our internal users are satisfied with the performance, but one particular user has had all sorts of performance related issues affecting their adoption and usage of the NFS shares.  From initial analysis, this seems to be a problem of loading (because it occurs during intensive reads/writes of large datasets, such as during rsync transfers or backup reads).

So far, I have not yet been able to establish whether the problem is due to the number of files or the volume of data, and I am now setting up test client nodes to try and reproduce the client's issues.

I googled around seeking information on tuning (kernel, NFS, OS, etc), but cannot seem to find any at all.   Please, can someone share some insight into how to go about trying to optimize the performance of ganesha.  Also, would appreciate any insight /suggestion on how best to benchmark the performance. 

Thanks in advance
_______________________________________________
Support mailing list -- support@lists.nfs-ganesha.org
To unsubscribe send an email to support-leave@lists.nfs-ganesha.org