Why Neo4j Dump Size Is So Small

[johnson_zhangc]

I'm using Neo4j Community Edition on RHEL, and load a graph with more than 1,000 nodes, and when I use neo4j-admin dump command, it says: Done: 34 files, 251.4MiB processed.

But when I check the final dump file, it's only 1.1MB. Does anybody know what is the problem?

Thanks in advance.

[dana_canzano]

@johnson.zhangc

what version of Neo4j?

For v4.x what is the size if data/dataases/ and data/transactions/?
For v3.x what is the size of data/graph.db?

if you run neo4j-admin load ..... does it properly read the dump file and recreate the database.

neo4j-admin dump does use GitHub - luben/zstd-jni: JNI binding for Zstd so as to compress the resultant dump.

[johnson_zhangc]

Here is the more information regarding on the dump and restore

I'm using neo4j 4.3.2

data/databases/neo4j database size is 1.5M
data/transactions/neo4j size is 250M

And if I run neo4j-admin load, after load, the database will become offine.

sudo neo4j-admin load --from=neo4j-210803.dump --force --verbose

Selecting JVM - Version:11.0.11, Name:OpenJDK 64-Bit Server VM, Vendor:Red Hat, Inc.
neo4j 4.3.2
VM Name: OpenJDK 64-Bit Server VM
VM Vendor: Red Hat, Inc.
VM Version: 11.0.11+9-LTS
JIT compiler: HotSpot 64-Bit Tiered Compilers
VM Arguments: [-XX:+UseG1GC, -XX:-OmitStackTraceInFastThrow, -XX:+AlwaysPreTouch, -XX:+UnlockExperimentalVMOptions, -XX:+TrustFinalNonStaticFields, -XX:+DisableExplicitGC, -XX:MaxInlineLevel=15, -XX:-UseBiasedLocking, -Djdk.nio.maxCachedBufferSize=262144, -Dio.netty.tryReflectionSetAccessible=true, -Djdk.tls.ephemeralDHKeySize=2048, -Djdk.tls.rejectClientInitiatedRenegotiation=true, -XX:FlightRecorderOptions=stackdepth=256, -XX:+UnlockDiagnosticVMOptions, -XX:+DebugNonSafepoints, -Dlog4j2.disable.jmx=true, -Dfile.encoding=UTF-8]
Done: 34 files, 251.4MiB processed.
The loaded database is not on the latest format (current:SF4.0.0, latest:SF4.3.0). Set dbms.allow_upgrade=true to enable migration.

[dana_canzano]

@johnson.zhangc

regarding the file size etc I am getting the same, for example

 ./neo4j-admin dump --database=july20 --to=/tmp
Selecting JVM - Version:11.0.7, Name:OpenJDK 64-Bit Server VM, Vendor:Ubuntu
Done: 70 files, 274.3MiB processed.
ls -lh /tmp/july20.dump
-rw-rw-r-- 1 neo4j neo4j 6.0M Aug  4 11:49 /tmp/july20.dump

so whereas the 'dump' command reports 70 files, 274.3MiB processed. the resultant dump file is only 6.0M.
And if I then se neo4j-admin load

./neo4j-admin load --database=july201 --verbose --from=/tmp/july20.dump --force
Selecting JVM - Version:11.0.7, Name:OpenJDK 64-Bit Server VM, Vendor:Ubuntu
neo4j 4.3.2
VM Name: OpenJDK 64-Bit Server VM
VM Vendor: Ubuntu
VM Version: 11.0.7+10-post-Ubuntu-2ubuntu219.10
JIT compiler: HotSpot 64-Bit Tiered Compilers
VM Arguments: [-XX:+UseG1GC, -XX:-OmitStackTraceInFastThrow, -XX:+AlwaysPreTouch, -XX:+UnlockExperimentalVMOptions, -XX:+TrustFinalNonStaticFields, -XX:+DisableExplicitGC, -XX:MaxInlineLevel=15, -XX:-UseBiasedLocking, -Djdk.nio.maxCachedBufferSize=262144, -Dio.netty.tryReflectionSetAccessible=true, -Djdk.tls.ephemeralDHKeySize=2048, -Djdk.tls.rejectClientInitiatedRenegotiation=true, -XX:FlightRecorderOptions=stackdepth=256, -XX:+UnlockDiagnosticVMOptions, -XX:+DebugNonSafepoints, -Dlog4j2.disable.jmx=true, -Dfile.encoding=UTF-8]
Done: 70 files, 274.3MiB processed.

and then start cypher-shell and run a create database july201; all appears as expected.

So the small size, maybe its simply a matter of compression. And I am yet to prove any failure as a result of its size

[johnson_zhangc]

@dana.canzano

Thank you very much Dana. I was doubting the dump size just because I never successfully load a dump file, every time it will bring the database to offline. Maybe this is the problem of community version?