which registry ticket storage ?

[Daniel CHARLOT]

Hello there,

We need to upgrade our CAS server V4 to V6 and add an loadbalancer F5.
We will do some HA with them and we wondering which is the product with the best performance for use ticket registry ?
Memcached, cassanadra, mongodb, couchbase redis or an other ?
I would like choose redis, because its seems works like a charm, but im not sure.

Thanks for your advices and your experiences return.

Daniel CHARLOT

[Ray Bon]

Daniel,

You could create a load test (maybe there is one already with cas) and check the various options.

We switched to hazelcast, for cas version 5, from ehcache. Our confg can handle several logins per second (I do not remember the exact numbers) for an extended period of time.

Ray

On Wed, 2021-03-17 at 15:50 +0100, Daniel CHARLOT wrote:

Notice: This message was sent from outside the University of Victoria email system. Please be cautious with links and sensitive information.

--

- Website: 

https://apereo.github.io/cas

- Gitter Chatroom: 

https://gitter.im/apereo/cas

- List Guidelines: 

https://goo.gl/1VRrw7

- Contributions: 

https://goo.gl/mh7qDG

---

You received this message because you are subscribed to the Google Groups "CAS Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 

cas-user+u...@apereo.org

.

To view this discussion on the web visit 

https://groups.google.com/a/apereo.org/d/msgid/cas-user/C22D6DFE-95B2-4D8E-92D3-15EFF6A17D1E%40unice.fr

.

-- 

Ray Bon

Programmer Analyst

Development Services, University Systems

2507218831 | CLE 019 | rb...@uvic.ca

I respectfully acknowledge that my place of work is located within the ancestral, traditional and unceded territory of the Songhees, Esquimalt and WSÁNEĆ Nations.

[Julien Gribonvald]

Hi,

I would recommand Redis, after memcached it's the fastest and more
efficient system for large number of datas. Other will works well but
you will win in effiency and server ressources comparing to other systems.

I don't recommand memcached only because we can't watch on datas without
modifying/refreshing it, the problem is that we can't have a good
overview on and on the datas.

Thanks

Julien Gribonvald

[Jérôme Rautureau]

Hi guys and girls,

at La Rochelle, we switch from galera-mariadb-3-nodes-cluster with jpa and CAS 4.2.x version to mongodb-replica-set-3-nodes and CAS 6.2.x.

We had a lot of issues with JPA (deadlocks) and now it's all about performance.

Only one issue solved by : https://github.com/apereo/cas/commit/9545d7aca56022e55a872c8a96597e7fd1656863 witch we have backported from 6.4 branch to 6.2.x branch.

I didn't test the other ticket registries.

--
- Website: https://apereo.github.io/cas
- Gitter Chatroom: https://gitter.im/apereo/cas
- List Guidelines: https://goo.gl/1VRrw7
- Contributions: https://goo.gl/mh7qDG
---
You received this message because you are subscribed to the Google Groups "CAS Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cas-user+u...@apereo.org.

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/9f0049ae-38fb-6637-9d17-ea7820b07e61%40recia.fr.

[King, Robert]

Have not tried these yet, but these load tests do exist on the website:

 

https://apereo.github.io/cas/6.3.x/high_availability/High-Availability-Performance-Testing.html

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/3b69ff5296b8614658aa7fc489beebab4f30ba5b.camel%40uvic.ca.

[Daniel CHARLOT]

Hi Robert, Jérome, Julien and Ray,

Many thanks for your responses. 

It was very usefull for me to have many feedback from your experiences.

 

I know where i go now :)

Have a nice day !

Daniel CHARLOT Architecte Système et Logiciels Tél : 04-92-07-67-07 Petit Valrose Avenue Joseph Valot 06108 Nice Cedex 2 univ-cotedazur.fr		DIRECTION DES SYSTÈMES D'INFORMATIONS Ingénierie et Administration Technique

Daniel CHARLOT Architecte Système et Logiciels Tél : 04-92-07-67-07 Petit Valrose Avenue Joseph Valot 06108 Nice Cedex 2 univ-cotedazur.fr		DIRECTION DES SYSTÈMES D'INFORMATIONS Ingénierie et Administration Technique

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/614936b2da564970ab1af11a0f1ac9c2%40mun.ca.

[Jérôme NENERT]

Le 17/03/2021 à 17:34, Julien Gribonvald a écrit :
> Hi,
>
Hi Julien,

> I would recommand Redis, after memcached it's the fastest and more
> efficient system for large number of datas. Other will works well but
> you will win in effiency and server ressources comparing to other
> systems.

Which type of Redis replication do you use in combination which CAS ?
Sentinel with redis.sentinel configuration keys or other type of
configuration ?

Jerome Nenert

[Julien Gribonvald]

Hi Jérôme,

We are deploying for several services the same conf: 3 serveurs with
redis + sentinel each one (We think also to deploy a redis farm in
place). One server is a master and the two other are slaves, we don't
use the persistant option nor the database. Also we use a HAproxy when
we can distinguish read and write access to share the load. All this
conf was defined mainly from this blog:
https://www.willandskill.se/en/setup-a-highly-available-redis-cluster-with-sentinel-and-haproxy/


Julien Gribonvald