memcached failover

[Niva Agmon]

>>Since it doesn't replicate, I would assume so, yes.
Memcached does replicate:

test-1:
16:37:51.675965 IP .53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [P.], seq 363627134:363627222, ack 465451673, win 270, options [nop,nop,TS val 640950434 ecr 365519181], length 88
16:37:51.676069 IP xxxtest-1.xxxxx.xxx.memcache > xxxtest-2.xxxxx.xxx.53469: Flags [P.], seq 1:34, ack 88, win 553, options [nop,nop,TS val 366875744 ecr 640950434], length 33
16:37:51.680348 IP xxxtest-2.xxxxx.xxx.53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [.], ack 34, win 270, options [nop,nop,TS val 640950439 ecr 366875744], length 0
16:37:51.681040 IP xxxtest-2.xxxxx.xxx.53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [P.], seq 88:200, ack 34, win 270, options [nop,nop,TS val 640950440 ecr 366875744], length 112
16:37:51.681195 IP xxxtest-1.xxxxx.xxx.memcache > xxxtest-2.xxxxx.xxx.53469: Flags [P.], seq 34:58, ack 200, win 553, options [nop,nop,TS val 366875749 ecr 640950440], length 24
...

test-2:
16:37:51.669026 IP xxxtest-2.xxxxx.xxx.53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [P.], seq 363627134:363627222, ack 465451673, win 270, options [nop,nop,TS val 640950434 ecr 365519181], length 88
16:37:51.673377 IP xxxtest-1.xxxxx.xxx.memcache > xxxtest-2.xxxxx.xxx.53469: Flags [P.], seq 1:34, ack 88, win 553, options [nop,nop,TS val 366875744 ecr 640950434], length 33
16:37:51.673416 IP xxxtest-2.xxxxx.xxx.53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [.], ack 34, win 270, options [nop,nop,TS val 640950439 ecr 366875744], length 0
16:37:51.674248 IP xxxtest-2.xxxxx.xxx.53469 > xxxtest-1.xxxxx.xxx.memcache: Flags [P.], seq 88:200, ack 34, win 270, options [nop,nop,TS val 640950440 ecr 366875744], length 112
16:37:51.676753 IP xxxtest-1.xxxxx.xxx.memcache > xxxtest-2.xxxxx.xxx.53469: Flags [P.], seq 34:58, ack 200, win 553, options [nop,nop,TS val 366875749 ecr 640950440], length 24



-----Original Message-----
From: users [mailto:users-bounces at shibboleth.net] On Behalf Of Cantor, Scott
Sent: Monday, August 1, 2016 2:56 PM
To: Shib Users <users at shibboleth.net>
Subject: RE: memcached failover

> > Because of CAS? If not, why?
> Yes, for failover for all SSO sessions, Shib (SAML) & CAS.

SSO sessions are maintained by the client automatically, it's the CAS piece that requires server-side storage at the moment. Hopefully we can get the ticket logic converted to avoid that requirement.

-- Scott

-- 
To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net