You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ilya Lantukh (JIRA)" <ji...@apache.org> on 2018/07/16 09:17:00 UTC

[jira] [Assigned] (IGNITE-9006) Add per-call tracing capabilities for IgniteCache#{get, getAll} methods

     [ https://issues.apache.org/jira/browse/IGNITE-9006?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ilya Lantukh reassigned IGNITE-9006:
------------------------------------

    Assignee: Ilya Lantukh

> Add per-call tracing capabilities for IgniteCache#{get, getAll} methods
> -----------------------------------------------------------------------
>
>                 Key: IGNITE-9006
>                 URL: https://issues.apache.org/jira/browse/IGNITE-9006
>             Project: Ignite
>          Issue Type: New Feature
>    Affects Versions: 2.6
>            Reporter: Alexey Goncharuk
>            Assignee: Ilya Lantukh
>            Priority: Major
>             Fix For: 2.7
>
>
> As an experimental feature to debug objects visibility issues, we should add an API method {{IgniteCache#withTrace()}} which will enable per-call API calls tracing.
> I think we can start with reads tracing, as a simplest change. We need to introduce an additional flag mask to {{NearSingleGetRequest}}, {{NearGetRequest}}, {{Near...Response}} to identify such requests, and when processing, print out all message processing path (send, receive), entry swap/unswap, versions and values read, {{GridCacheEntryRemovedException}}s encountered, etc. For transactional caches, I think it also makes sense to print out all pending transactions that touch the key being read.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)