[
https://issues.apache.org/jira/browse/OAK-6071?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16010165#comment-16010165
]
Michael Dürig commented on OAK-6071:
------------------------------------
I'm not sure whether his is actually helpful. In my experience Warnings in the
logs tend to get mostly ignored. To that respect OAK-1454 didn't caused any
useful action to my knowledge.
Also the patch would create a lot of noise. In a recent analysis on a
production repository I found
* 32.0E+6 strings s with 128 chars < s < 1k chars,
* 450.3E+3 strings s with 1k chars < s < 16k chars,
* 8.3E+3 strings s with 16k chars s < 1M chars and
* 3.0E+0 strings s with s > 1M chars.
Can't we collect this kind of information on the fly (i.e. through a query or
tooling)? Another approach would be to implement a commit hook that gathers
such kind of information/statistics.
> Flag string property of large size
> ----------------------------------
>
> Key: OAK-6071
> URL: https://issues.apache.org/jira/browse/OAK-6071
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: jcr
> Reporter: Chetan Mehrotra
> Assignee: Chetan Mehrotra
> Priority: Minor
> Fix For: 1.8
>
> Attachments: OAK-6071-v1.patch
>
>
> At times users store large text values as String property which causes issues
> later in form of higher resource consumption of caches, higher repository
> size etc.
> It would be good to have a validator which can flag such strings at time of
> commit. The size limit can be the one we use for in lined binary like 1-2KB
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
