good point Mav. This is a long discussion around the ticket registry design and 
I think I created a Jira a long time ago gor this. It might be enough to expose 
the ticket timestamp in an own attribute of the registry so a deserialization 
of all tickets in the cleaner wouldn't be needed. Disabling SLO is not an 
option in our case, as you do not know when a logout occurs and users expect to 
be logged out from all services on SLO.

Sent while mobile.

Am 18.12.2014 um 16:12 schrieb Marvin Addison <[email protected]>:

>> When I got a heap dump I was shocked to see that there were tickets in there 
>> which took close to 10Mb of memory (deserialized). After investigation this 
>> was pretty much all allocated in the 'services' map that is stored in the 
>> TGT.
> 
> This is a known issue and a great example of why we need to tighten up the 
> storage model in future versions of CAS. We hold on to way more data than is 
> needed to track accessed services (needed for logout). That said, despite 
> improvements and optimizations, it's not unreasonable to expect problems from 
> an SSO session that tracks service sessions for 90 days. We might want to 
> expose a configuration knob to turn off session tracking since the only case 
> for that is SLO, which is meaningless over a span that long. It seems like 
> you pretty much came to a similar conclusion on your own.
> 
> You might reframe your problem as a design suggestion and discuss further on 
> cas-dev.
> 
> M
> 
> -- 
> 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

Reply via email to