Closed WAG10 closed 1 year ago
Hi @gregwood-db , thanks for the fix. Unfortunately i still get the namespace error.(I cloned the repo from manin and manually made the changes as per fix from your commit)
cloudsa@mylinuxserver:~/export_dir_4$ python3 /home/cloudsa/migrate_repo_fix2/migrate/import_db.py --azure --profile newWS --metastore --skip-failed
Note: running import_db.py directly is not recommended. Please use migration_pipeline.py
Traceback (most recent call last):
File "/home/cloudsa/migrate_repo_fix2/migrate/import_db.py", line 290, in
@WAG10 is there a reason to use import_db.py
instead of migration_pipeline.py
? We're mostly testing against the latter for new features/flags. I'll have to go through all of the flags to see which are currently missing from the import and export pipelines.
hi @gregwood-db , We are using some terraform utilities for exporting other databricks components as per internal architecture. We are using this migrate repo to import/export hive metastore alone hence using import_db.py. If you have some syntax or way which I can use with migration_pipeline.py to import/export hive metastore alone, I will be more than happy to do :) I tried myself couple of times but i was getting big bunch of errors hence to decided to switch import_db.py.
Did you try migration_pipeline.py --import-pipeline --keep-tasks metastore
? That should run only the metastore task for the import pipeline.
@gregwood-db , Many thanks for the lead. I tried the below command for export first.
_cloudsa@mylinuxserver:~/migrate_repo_fix2/migrate$ python3 migrationpipeline.py --export-pipeline --profile oldWS --azure --keep-tasks metastore
This exported me the hms metadata.
Then i tried running import using below command.
_cloudsa@mylinuxserver:~/migrate_repo_fix2/migrate$ python3 migrationpipeline.py --import-pipeline --profile newWS --azure --keep-tasks metastore
But here i got below error
_Using the session id: M20230906094058
Traceback (most recent call last):
File "migration_pipeline.py", line 378, in
This was because export which i ran generated different session id folder and import was trying to find the /source_info.txt in the session id folder which got created for import. Any way to solve this issue?
Yes- please specify --session <session-id>
where session-id is the session token you got from the export job. This will use the existing export directory to import to the new WS.
Closing this- please let us know if you're still having issues.
Hi Team we are using below command tp import the hive metarstore from primary to secondary workspace. When we run the command we get the error as : Attribute Error: 'Namespace' object has no attribute 'timeout' We commented out the occurrence of the timeout and then it worked but this seems not a ideal way to handle. Requesting your inputs on this.
Command : _python3 /home/cloudsa/migrate_repo_fix/migrate/importdb.py --azure --profile newWS --metastore --skip-failed
Error: _Note: running import_db.py directly is not recommended. Please use migration_pipeline.py Traceback (most recent call last): File "/home/cloudsa/migrate_repo_fix/migrate/import_db.py", line 290, in
main()
File "/home/cloudsa/migrate_repo_fix/migrate/import_db.py", line 27, in main
client_config = build_client_config(args.profile, url, token, args)
File "/home/cloudsa/migrate_repo_fix/migrate/dbclient/parser.py", line 432, in build_clientconfig
'timeout': args.timeout,
AttributeError: 'Namespace' object has no attribute 'timeout'