You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@geode.apache.org by "Geode Integration (Jira)" <ji...@apache.org> on 2021/12/08 01:09:00 UTC

[jira] [Commented] (GEODE-9849) validate serialized blobs when they are stored on a server

    [ https://issues.apache.org/jira/browse/GEODE-9849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17454908#comment-17454908 ] 

Geode Integration commented on GEODE-9849:
------------------------------------------

Seen in [acceptance-test-openjdk11 #42|https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/acceptance-test-openjdk11/builds/42] ... see [test results|http://files.apachegeode-ci.info/builds/apache-develop-main/1.15.0-build.0717/test-results/acceptanceTest/1638913188/] or download [artifacts|http://files.apachegeode-ci.info/builds/apache-develop-main/1.15.0-build.0717/test-artifacts/1638913188/acceptancetestfiles-openjdk11-1.15.0-build.0717.tgz].

> validate serialized blobs when they are stored on a server
> ----------------------------------------------------------
>
>                 Key: GEODE-9849
>                 URL: https://issues.apache.org/jira/browse/GEODE-9849
>             Project: Geode
>          Issue Type: Improvement
>          Components: client/server, core, serialization
>            Reporter: Darrel Schneider
>            Priority: Major
>
> When a client sends a value to a server, geode try to not deserialize it. It instead just takes the bytes sent to it from the client and stores them on the server. But is some very rare instances the blob may be corrupt. It would be nice if the server could validate the blob BEFORE storing it (and distributing it to other servers and writing to a disk store and sending it over a gateway).
> How much validation geode can do without impacting performance and without requiring the server to be able to fully deserialize it not clear. But one validation it could definitely do is to check the first byte of the blob and make sure it is a valid DSCODE. This would detect some of the corruptions we have seen in the past (i.e. "unknown header byte" issues) and be a very efficient check.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)