[
https://issues.apache.org/jira/browse/OAK-12114?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18061652#comment-18061652
]
Rishabh Daim commented on OAK-12114:
------------------------------------
More on the BSON Wired Size vs Storage Size:
{quote}What the method creates
char[] chars = new char[1024 * 1024]; // 1M chars
Arrays.fill(chars, '0'); // all identical: "00000000..."
String content = new String(chars);
BSON Wire Size vs Storage Size
These are two separate things:
1. BSON Wire Size (~1MB)
The '0' character is ASCII (1 byte in UTF-8), so the string is exactly ~1MB
as BSON on the wire. This is what matters for the
BsonMaximumSizeExceededException — the 16MB limit is enforced on the
uncompressed BSON
payload before it reaches MongoDB.
2. Storage Size (near zero)
MongoDB's WiredTiger engine (default since 3.x) uses Snappy compression on
disk. A string of 1 million identical characters is maximally compressible —
Snappy would reduce it to just a few bytes using run-length
encoding.{quote}
> MongoDBExceptionTest uses payloads that compress well
> -----------------------------------------------------
>
> Key: OAK-12114
> URL: https://issues.apache.org/jira/browse/OAK-12114
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: mongomk
> Reporter: Julian Reschke
> Priority: Major
> Fix For: 2.0.0
>
>
> ...containing only zeros, thus defeating the purpose of testing large
> payloads.
> (actually, not testing it at all)
> With this change:
> {code}
> diff --git
> a/oak-store-document/src/test/java/org/apache/jackrabbit/oak/plugins/document/mongo/MongoDBExceptionTest.java
>
> b/oak-store-document/src/test/java/org/apache/jackrabbit/oak/plugins/document/mongo/MongoDBExceptionTest.java
> index d6e7cb264e..934af36627 100644
> ---
> a/oak-store-document/src/test/java/org/apache/jackrabbit/oak/plugins/document/mongo/MongoDBExceptionTest.java
> +++
> b/oak-store-document/src/test/java/org/apache/jackrabbit/oak/plugins/document/mongo/MongoDBExceptionTest.java
> @@ -36,7 +36,6 @@ import org.junit.Test;
> import org.slf4j.event.Level;
> import java.util.ArrayList;
> -import java.util.Arrays;
> import java.util.List;
> import static java.util.Collections.singletonList;
> @@ -333,9 +332,5 @@ public class MongoDBExceptionTest {
> // RED ALERT: OAK-12114
> private String create1MBContent() {
> - char[] chars = new char[1024 * 1024];
> - Arrays.fill(chars, '0');
> - String content = new String(chars);
> - return content;
> - }
> + return RandomStringUtils.secure().next(1024 * 1024); }
> }
> {code}
> we get
> {code}
> ERROR] Failures:
> [ERROR] MongoDBExceptionTest.createOrUpdate16MBDoc:156
> Expected: a string containing "Document to upsert is larger than 16777216"
> but: was "Document size of 37633707 is larger than maximum of 16793600.
> [/foo]"
> [ERROR] MongoDBExceptionTest.findAndUpdate16MBDoc:304
> Expected: a string containing "Resulting document after update is larger than
> 16777216"
> but: was "Payload document size is larger than maximum of 16793600.
> [/foo]"
> [ERROR] MongoDBExceptionTest.multiCreateOrUpdate16MBDoc:263
> Expected: a string containing "Resulting document after update is larger than
> 16777216"
> but: was "Payload document size is larger than maximum of 16793600.
> [/test]"
> [ERROR] MongoDBExceptionTest.update16MBDoc:231
> Expected: a string containing "Resulting document after update is larger than
> 16777216"
> but: was "Payload document size is larger than maximum of 16793600.
> [/foo]"
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
