#
261055 |
|
22-Jan-2014 |
mav |
MFC r260229, r260258, r260367, r260390, r260459, r260648: Rework NFS Duplicate Request Cache cleanup logic.
- Introduce additional hash to group requests by hash of sockref. This allows to process TCP acknowledgements without looping though all the cache, and as result allows to do it every time. - Indroduce additional callbacks to notify application layer about sockets disconnection. Without this last few requests processed just before socket disconnection never processed their ACKs and stuck in cache for many hours. - Implement transport-specific method for tracking reply acknowledgements. New implementation does not cross multiple stack layers to get the data and does not have race conditions that previously made some requests stuck in cache. This could be done more efficiently at sockbuf layer, but that would broke some KBIs, while I don't know other consumers for it aside NFS. - Instead of traversing all DRC twice per request, run cleaning only once per request, and except in some conditions traverse only single hash slot at a time.
Together this limits NFS DRC growth only to situations of real connectivity problems. If network is working well, and so all replies are acknowledged, cache remains almost empty even after hours of heavy load. Without this change on the same test cache was growing to many thousand requests even with perfectly working local network.
As another result this reduces CPU time spent on the DRC handling during SPEC NFS benchmark from about 10% to 0.5%.
Sponsored by: iXsystems, Inc.
|
#
254337 |
|
14-Aug-2013 |
rmacklem |
Fix several performance related issues in the new NFS server's DRC for NFS over TCP. - Increase the size of the hash tables. - Create a separate mutex for each hash list of the TCP hash table. - Single thread the code that deletes stale cache entries. - Add a tunable called vfs.nfsd.tcphighwater, which can be increased to allow the cache to grow larger, avoiding the overhead of frequent scans to delete stale cache entries. (The default value will result in frequent scans to delete stale cache entries, analagous to what the pre-patched code does.) - Add a tunable called vfs.nfsd.cachetcp that can be used to disable DRC caching for NFS over TCP, since the old NFS server didn't DRC cache TCP. It also adjusts the size of nfsrc_floodlevel dynamically, so that it is always greater than vfs.nfsd.tcphighwater.
For UDP the algorithm remains the same as the pre-patched code, but the tunable vfs.nfsd.udphighwater can be used to allow the cache to grow larger and reduce the overhead caused by frequent scans for stale entries. UDP also uses a larger hash table size than the pre-patched code.
Reported by: wollman Tested by: wollman (earlier version of patch) Submitted by: ivoras (earlier patch) Reviewed by: jhb (earlier version of patch) MFC after: 1 month
|
#
191783 |
|
04-May-2009 |
rmacklem |
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building.
Approved by: kib (mentor)
|