seaweedfs: Can't upload on s3 bucket.

Hello!

Describe the bug I’ve run seaweed with s3 (3 masters, 3 volumes, 3 filers on the same servers). It works good. Some time later I have a problem. Create s3-bucket with s3cmd: s3cmd mb s3://test It succefully created. That I tried to upload some file. s3cmd put file s3://test I’ve got error: WARNING: Upload failed: /file (500 (InternalError): We encountered an internal error, please try again.) But on older bucket I can upload files.

System Setup

List the command line to start “weed master”, “weed volume”, “weed filer”, “weed s3”, “weed mount”. /opt/seaweedfs/weed master -mdir=/data/seaweedfs/master -peers=10.214.3.19:9333,10.214.3.16:9333,10.214.3.17:9333 -volumeSizeLimitMB 1024 /opt/seaweedfs/weed volume -mserver=10.214.3.19:9333,10.214.3.16:9333,10.214.3.17:9333 -dir=/data/seaweedfs/volume -dataCenter dc1 -rack rack1 -ip=10.214.3.16 -max=0 /opt/seaweedfs/weed filer -master 10.214.3.19:9333,10.214.3.16:9333,10.214.3.17:9333 -s3 -s3.config /etc/seaweedfs/s3.config.json -s3.domainName example.com -s3.port 80
OS version CentOS Linux release 7.8.2003 (Core)
output of weed version version 30GB 2.11 98827d6 linux amd64
if using filer, show the content of filer.toml Here is part of filer.toml

[postgres] # or cockroachdb
# CREATE TABLE IF NOT EXISTS filemeta (
#   dirhash     BIGINT,
#   name        VARCHAR(65535),
#   directory   VARCHAR(65535),
#   meta        bytea,
#   PRIMARY KEY (dirhash, name)
# );
enabled = "True"
hostname = "10.214.3.19"
port = 5432
username = "seaweedfs_user"
password = "SECRET_PASSWORD"
database = "seaweedfs_db"              # create or use an existing database
sslmode = "disable"
connection_max_idle = 100
connection_max_open = 100

Expected behavior I can upload file to s3.

Additional context Also I have the same problem with another s3 tools (aws, s3 sync)

About this issue

Original URL
State: closed
Created 4 years ago
Comments: 16 (12 by maintainers)

Most upvoted comments

Each volume is configured to be 1GB. -volumeSizeLimitMB 1024 Each bucket will create 7 volumes by default.

The folder /data/seaweedfs/volume seems having the number of volumes multiply 1GB close to the disk limit.

You can reduce to -volumeSizeLimitMB 512.

If you are using the git master branch, to be released in 2.12, there is a more flexible configuration:

If you have a lot of buckets to add, you can configure the per bucket storage this way in weed shell:

> fs.configure -locationPrefix=/buckets/ -volumeGrowthCount=1 -apply

This will add 1 physical volume when existing volumes are full. If using replication, you will need to add more volumes.

See https://github.com/chrislusf/seaweedfs/wiki/Path-Specific-Configuration

chrislusf on Nov 20, 2020