Backup size ~twice as large as index size - normal behavior?

Elastic Stack Elasticsearch
1

Hi,

i just made a snapshot of indices, with a combined size of 8.5GB.

The snapshot is about 16GB in size.

Is this normal for snapshots, or did I miss a setting/compression?

    PUT /_snapshot/backup1
{
    "type": "fs",
    "settings": {
        "location": "/mnt/nas/es_backup/backup1",
        "compress": true
    }
}

PUT /_snapshot/backup1/snapshot_2015_10_22
{
  "indices": "*10.22"
}

[root@prod00elastic11 backup1]# pwd
/mnt/nas/es_backup/backup1
[root@prod00elastic11 backup1]# du -sh *
1.0M    index
16G     indices
1.0M    metadata-snapshot_2015_10_22
1.0M    snapshot-snapshot_2015_10_22

I had a total of 50 indices, each with 5 shards and 1 replica if thats relevant

2

@Pablo_Musa (hope that I'm pinging the right pablo here), you were right after all. The file system on my nfs messes something up. I guess I will have to look what I did wrong. This is definitely not an Elasticsearch issue :slight_smile:

###first test:
size on elasticsearch /data directory
[root@prod00elastic07 wintelu]# du -sh test1.file
216K test1.file

copy file to nfs backup path

size on elasticsearch /nfs directory

[root@prod00elastic07 backup1]# du -sh *
1.0M test1.file

five fold increase in size!

###second test:
size on elasticsearch /data directory
[root@prod00elastic07 wintelu]# du -sh test.file
3.8M test.file

copy file to nfs backup path

size on elasticsearch /nfs directory
[root@prod00elastic07 backup1]# du -sh test.file
6.0M test.file

about 50% larger than before

3

That is great to know Luca :wink:

Good luck!