[infinispan-issues] [JBoss JIRA] (ISPN-2491) Order of locks in optimistic locking is not strict

[Radim Vansa (JIRA)]

     [ https://issues.jboss.org/browse/ISPN-2491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Radim Vansa closed ISPN-2491.
-----------------------------
    Resolution: Out of Date


Locking is not ordered by hashcodes anymore. The locks on each primary are requested in a synchronized block, introducing order on the requests. The locks are fair, and therefore two transactions sharing a set of keys will have the subset served in the same order.

> Order of locks in optimistic locking is not strict
> --------------------------------------------------
>
>                 Key: ISPN-2491
>                 URL: https://issues.jboss.org/browse/ISPN-2491
>             Project: Infinispan
>          Issue Type: Quality Risk
>          Components: Transactions
>    Affects Versions: 5.1.8.Final, 5.2.0.Beta3
>            Reporter: Radim Vansa
>            Priority: Minor
>
> In OptimisticLockingInterceptor, the keys are ordered according to their hash. However, the hashes can still collide, which may result in a deadlock if two keys with identical hash (only 32-bit) are sorted to different order. We should try to check if the keys are Comparable or let user provide some comparator class in config, and use the compare of hash only as the last resort.
> In all cases, a warning should be emitted if the compare operation had non-strict result.



--
This message was sent by Atlassian JIRA
(v7.5.0#75005)