Hi, have some issues with lakectl local: i created a directory where I want to clone from remote to local. I have 3 branches created correctly. datalakeadmin@prj-sm-etm-dev-datalake-eval:~/ETM/lakefs/project_folder/lakefs/clone_target_dir$ lakectl local clone lakefs://datalake/main/ directory '/home/datalakeadmin/ETM/lakefs/project_folder/lakefs/clone_target_dir' exists and is not empty The issue is when I run ls -l in my working local directory it returns no objects. It seems the clone 'worked' yet retrieved no data files. Also when I enter lakefs local pull It works but no actual files are retrieved from remote. lakefs local list shows a valid linking with the server: lakectl local list: +-----------+-------------------------+------------------------------------------------------------------+ | DIRECTORY | REMOTE URI | SYNCED COMMIT | +-----------+-------------------------+------------------------------------------------------------------+ | . | lakefs://datalake/main/ | 4b846d749af8956bd1082f930d58258ff3128e4d0f00c0897b7794f4fafc7d35 | +-----------+-------------------------+------------------------------------------------------------------+

@Patrick Glynn Sorry to hear you're having issues with lakectl local. Can you tell me what you see when you run the command:

lakectl local status ~/ETM/lakefs/project_folder/lakefs/clone_target_dir

I get this output: lakectl local status ~/ETM/lakefs/project_folder/lakefs/clone_target_dir diff 'local:///home/datalakeadmin/ETM/lakefs/project_folder/lakefs/clone_target_dir' <--> 'lakefs://datalake/4b846d749af8956bd1082f930d58258ff3128e4d0f00c0897b7794f4fafc7d35/'... diff 'lakefs://datalake/4b846d749af8956bd1082f930d58258ff3128e4d0f00c0897b7794f4fafc7d35/' <--> 'lakefs://datalake/main/'... No diff found. I am not sure whether if the config.yaml is not linked correclty to the server instance or if I have misconfigured something minor

Do you by chance have more than one lakeFS server up in your env? Could lakectl be configured to a endpoint where repo and main branch are empty?

do I use lakefs diff to check that?

You can do that via the web UI or via

lakectl log

command

I ran the command and I can confirm

4b846d749af8956bd1082f930d58258ff3128e4d0f00c0897b7794f4fafc7d35

is the top commit

Can you run

lakectl fs ls <lakefs://datalake/main/>

I ran the command, it show the objects! 🙂

No, object should have been downloaded and you should have seen them

lakeFS version: 1.20.0

OK - I think I have a suspicion

returns nothing or executes and proposes a new cursor line

That's very strange - can you tell me a little bit about your system

its a Ubuntu linux system that is cmdline interface only. I am connecting to it via ssh Putty

Also can you run the command

ls -la

returns objects: total 12 drwxrwxr-x 2 datalakeadmin datalakeadmin 4096 Apr 16 07:36 . drwxrwxr-x 5 datalakeadmin datalakeadmin 4096 Apr 17 08:00 .. -rw-r--r-- 1 datalakeadmin datalakeadmin 124 Apr 17 08:25 .lakefs_ref.yaml

is that your clone directory?

yes its my clone directory

Then I want you to run:

lakectl local clone <lakefs://datalake/main/> test

it worked

I'm not exactly sure what happened. Perhaps you ran

lakectl local init

on that folder before you tried to clone? Basically we do not allow to perform clone on a directory which is not empty. Clone is an operation that consist of

init

and then

pull

. The init command is the one that creates the ref file which indicate to lakectl that this is a data directory. The extra parameter (

test

) is an optional argument telling the clone command what is that path to use. If not provided it uses the current working directory. If the provided path doesn't exist we try to create it.

I think yes I executed lakect local init before

I'm not sure I understand the question