ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aleksey Plekhanov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (IGNITE-5714) Implementation of suspend/resume for pessimistic transactions
Date Mon, 15 Apr 2019 12:25:00 GMT

    [ https://issues.apache.org/jira/browse/IGNITE-5714?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16817923#comment-16817923

Aleksey Plekhanov commented on IGNITE-5714:

[~agoncharuk], as already described by [~Alexey Kuznetsov] we can't remove thread ID from
MVCC candidate because it's needed for explicit locks (IgniteCache#lock, IgniteCache#lockAll)
when an explicit transaction is not started. In this case, the lock is held by a thread,
not by a transaction.

But I think we can get rid of thread id usage in MVCC candidate for all cases when the lock
is held by a transaction. And more carefully split the logic of lock checking for locks held
by threads and by transactions (for example remove methods like {{lockedLocallyByIdOrThread).}} In
this case, there is no thread id update needed when thread for the transaction is switched.

What do you think about such solution?

> Implementation of suspend/resume for pessimistic transactions
> -------------------------------------------------------------
>                 Key: IGNITE-5714
>                 URL: https://issues.apache.org/jira/browse/IGNITE-5714
>             Project: Ignite
>          Issue Type: Sub-task
>          Components: general
>            Reporter: Alexey Kuznetsov
>            Assignee: Aleksey Plekhanov
>            Priority: Major
>              Labels: iep-34
> Support transaction suspend()\resume() operations for pessimistic transactions. Resume
can be called in another thread.
>    _+But there is a problem+_: Imagine, we started pessimistic transaction in thread
T1 and then perform put operation, which leads to sending GridDistributedLockRequest to another
node. Lock request contains thread id of the transaction. Then we call suspend, resume in
another thread and we also must send messages to other nodes to change thread id. 
> It seems complicated task.It’s better to get rid of sending thread id to the nodes.
> We can use transaction xid on other nodes instead of thread id. Xid is sent to nodes
in GridDistributedLockRequest#nearXidVer
>    _+Proposed solution+_ : On remote nodes instead of thread id of near transaction GridDistributedLockRequest#threadId
use its xid GridDistributedLockRequest#nearXidVer.
> Remove usages of near transaction's thread id on remote nodes.

This message was sent by Atlassian JIRA

View raw message