[infinispan-issues] [JBoss JIRA] (ISPN-5298) HotRod millisecond precision for lifespan/maxidle

[Ion Savin (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-5298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13050796#comment-13050796 ] 

Ion Savin commented on ISPN-5298:
---------------------------------

If I understand correctly how expiration is supposed to work we should always be able to satisfy this condition (sync put) regardless of how small the duration/unit is:

put(k, v, lifespan, unit)
sleep(lifespan, unit)
assert(null, get(k))

If I recall correctly expiration is checked on access. So if the clocks are in sync with delta < unit I think it should work even for multiple-host setups.

Separate fields are a good option for large(r) durations with are not multiple of a second (e.g. 29day+23h+59m+59s+999ms). To reuse the existing fields their type would need to change to vLong to hold this duration in ns.

> HotRod millisecond precision for lifespan/maxidle
> -------------------------------------------------
>
>                 Key: ISPN-5298
>                 URL: https://issues.jboss.org/browse/ISPN-5298
>             Project: Infinispan
>          Issue Type: Feature Request
>          Components: Remote Protocols
>    Affects Versions: 7.2.0.Alpha1
>            Reporter: Ion Savin
>
> Currently lifespan/maxidle precision is SECONDS.
> JSR-107 requires MILLISECONDS precision.



--
This message was sent by Atlassian JIRA
(v6.3.11#6341)