As Andrew pointed out, memcache is generally deployed a large number of random servers that purely serve as caches. Thus, a client will usually only put the object on one of the servers, and the server is generally chosen by hashing the key.
Repcache will replicate information between two repcache instances that are aware of each other. This is independent of the memcached clients. On Tue, Jun 30, 2009 at 10:50 AM, Andrew Feller <[email protected]> wrote: > Andrew, > > I asked the same thing of Scott and apparently the memcached client will > generate a hash from the data to be stored and will use that to determine > which of the configured servers to send the information to. We are using > the memcached registry too, however I haven’t sat down and dug into the > memcached client code to figure out either, but that would be my > recommendation if you want to know what is going on. > > HTH, > A- > > > On 6/30/09 9:37 AM, "Andrew Tillinghast" <[email protected]> wrote: > > Can I get a little more clarification on memcache/repcache and CAS 3.3.3. > > MemCacheTicketRegistry constructor "0" is a list of hosts but the notes > specify that memcache doesn't provide replication. If the data isn't > replicated between servers then why a list of hosts? > > Repcache has very little documentation in the wiki, how would I add > repcache to the CAS 3.3.3 memcache and what would it actually add over and > above the 3.3.3 memcache? > > > -Andrew > > > On Jun 27, 2009, at 10:54 PM, Scott Battaglia wrote: > > On Fri, Jun 26, 2009 at 9:14 PM, venu.alla <[email protected]> wrote: > > hi, > I heard few schools have implemented repcache as CAS ticket registry. > > I would like to get some stats on : > 1. What are the hardware specs for CAS app > > > We have CAS and memcached running on two Sun T5120's (for redundancy). > > > 2. how much mem has been allotted to memcache. > > > I think we have one gig allocated. > > > 3. what is the user population size or usage in terms of TGT,ST count. > > > We have 50,000+ students/staff. I don't have an average TGT/ST count. > > > 4. How long did it take to re-sycn with the other node upon startup after > crash? and any other scenarios. > > > Its pretty fast, but I never actually calculated. > > -Scott > > > > > We are evaluating using memcache with repcache patch for ticketReg and > want to include other data as test cycles are short. > > Venu > > -- > You are currently subscribed to [email protected] as: > [email protected] > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user > > > > > > > -- > Andrew Feller, Analyst > LSU University Information Services > 200 Frey Computing Services Center > Baton Rouge, LA 70803 > Office: 225.578.3737 > Fax: 225.578.6400 > > -- > You are currently subscribed to [email protected] as: > [email protected] > > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user > > -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-user
