[Nbd] Scalability.
- To: nbd-general@lists.sourceforge.net
- Subject: [Nbd] Scalability.
- From: Dean Montgomery <dmonty@...954...>
- Date: Wed, 2 Nov 2011 10:41:08 -0700 (PDT)
- Message-id: <1854347818.1393871320255667948.JavaMail.root@...955...>
- In-reply-to: <1066952835.1392771320255509504.JavaMail.root@...955...>
I would like to know the scalability of NBD as root file system for handling 200 diskless clients on one server. We can currently handle 200 diskless clients on an NFS root file system on one server.
I just tested squashfs exported over NBD on a single diskless client. For a single client nbd/squashfs is a significant speed improvement in logging into KDE and launching firefox, openoffice, gimp, etc.
However on the server side I noticed that nbd-server is single threaded as opposed to multi-threaded NFS. Will the NBD server be able to handle 200 clients?
If NBD/squashfs is successful I will roll it out to about 3600 diskless clients at about 52 schools/sites.
--
Dean Montgomery
Network Support Tech./Programmer
http://dmonty.blogs.sd73.bc.ca/
dmonty@...954...
School District #73
Reply to: