version: v3.00.3 2025.05.15 LINUX x86_64
I'm running a single-node DolphinDB (Community Edition) instance, and I recently needed to migrate the entire data directory from /mnt/data2/DolphinDB to /mnt/data3/DolphinDB due to low disk space.
Here’s what I did:
# Stopped DolphinDB:
sudo systemctl stop dolphindb.service
# Verified the process was fully stopped:
ps aux | grep dolphindb
# Copied the entire directory using rsync:
sudo rsync -av --progress /mnt/data2/DolphinDB/ /mnt/data3/DolphinDB/
# Updated my systemd service file (/etc/systemd/system/dolphindb.service) by
# changing WorkingDirectory from /mnt/data2/DolphinDB/server to /mnt/data3/DolphinDB/server.
# Reloaded and restarted the service:
sudo systemctl daemon-reload
sudo systemctl start dolphindb.service
The server starts successfully, and newly created tables work fine. However, existing DFS tables (especially TSDB tables) are now corrupted in metadata. For example, when I try to drop an old table:
dropTable("dfs://forward_return", "zombie_close_neutralized")
I get this error:
Storage can't find volume for path [/mnt/data2/DolphinDB/server/local8848/storage/CHUNKS/forward_return/201001M_201101M/kTcm]
It seems the metadata still references the old path (/mnt/data2/...) even though all files have been moved to /mnt/data3/....
As a side question — for a single physical machine, would deploying a 2-node cluster (Community Edition) provide better performance than a single-node setup?
Any advice on safe migration practices or metadata recovery would be greatly appreciated!
