Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Hi Guys, We Consider Using Trains Experiment Manager And I Wonder Is There A Way To Backup All Experiment Data? In Case The Backend Machine Will Die And The Data Will Die With It.

Hi guys,
We consider using Trains Experiment Manager and I wonder Is there a way to backup all experiment data? in case the backend machine will die and the data will die with it.

Posted 4 years ago
Votes Newest

Answers 3

So you suggest that if I backup /opt/trains (assuming I use the docker-compose option on linux) I will be able to recover from it. Thanks for the fast reply

Posted 4 years ago

Indeed 🙂
Happy to help!

Posted 4 years ago

Hi AstonishingSeaturtle47 ,
Absolutely! 🙂
Whether running Trains Server locally (using docker-compose , for example) or in the cloud (using one of our AWS AMIs), you can always perform regular backups of the data directory (where all experiment data and metadata is stored).
See https://allegro.ai/docs/deploying_trains/trains_deploy_overview/ option #3 for information on how to deploy your own Trains Server, and https://allegro.ai/docs/deploying_trains/trains_server_upgrading/ which includes instructions on how to backup your data (under Data directory backup and restore example commands ).

Posted 4 years ago