> If I understand correctly, enabling fd caching caches the open fds and does > not close it until the belonging cache inode is evicted, which is quite > beneficial to NFSv3 protocol that may open/close on every read/write > without fd caching. (question: are there any chance that fds are closed?) > > Now, I want to limit the number of fds like, to 100. > > Is it possible to do this directly with some settings? > or is there some workaround to emulate that? > > I am seeing the wiki page and FD_*_Percent parameters may be what I am > looking for. > But I have no clear conclusion yet. > > https://github.com/nfs-ganesha/nfs-ganesha/wiki/CacheInodeParams
Yes, those FD parameters will manage the number of open fds. I'm not sure it prevents opening, but will apply back pressure to close fds above the high water mark. This is an area Ganesha really needs improvement, and 2.4.0 actually increases the number of fd opened since one is opened for every NFS v4 open and lock stateid, and for each NLM (NFS v3) lock and share owner (though only FSAL_VFS and FSAL_GPFS use actual "file descriptors" as returned by the open system call. FSAL_GLUSTER and FSAL_CEPH use file descriptor analogs. ulimit -n can also set a hard limit (though note that Ganesha does use some file descriptors for the log file and such). Frank --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ Nfs-ganesha-devel mailing list Nfs-ganesha-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs-ganesha-devel
