Dataiku DSS upgrade
Hi,
I want to upgrade from 5.1.6 to 10.
As far as I understand from this documentation ( https://doc.dataiku.com/dss/latest/release_notes/10.0.html#migration-paths-to-dss-10-0 ) the upgrade can be done directly without jumping from version to version. Can you please confirm my understanding?
Also, I understand that I only need to:
1. Stop DSS:
DATA_DIR/bin/dss stop
2. Download the new version:
wget https://cdn.downloads.dataiku.com/public/dss/10.0.2/dataiku-dss-10.0.2.tar.gz
3. Untar the file:
cd SOMEDIR
tar xzf /PATH/TO/dataiku-dss-NEWVERSION.tar.gz
4. Perform the upgrade:
dataiku-dss-NEWVERSION/installer.sh -d DATA_DIR -u
5. Start DSS
DATA_DIR/bin/dss start
Since I already have data, does this mean the installer will automatically 'convert' the existing data to work with the new DSS version and will also install all the dependencies that are needed?
Can you share some hints, best practices, etc? Do I need to pay special attention to something?
Thank you!
Operating system used: Linux (Ubuntu)
Best Answers
-
Sergey Dataiker, Dataiku DSS Core Designer, Dataiku DSS & SQL, Dataiku DSS Core Concepts Posts: 365 DataikerOptionsHi @darkmac
Yes, your understanding is correct. You can upgrade directly from DSS5.1 to DSS10. Just make sure you create a full backup of the DSS data directory before performing any upgrade:
https://doc.dataiku.com/dss/latest/operations/backups.html#full-backup
Please also don't forget about pre-start and post-start upgrade steps:
That should be all.
-
Sergey Dataiker, Dataiku DSS Core Designer, Dataiku DSS & SQL, Dataiku DSS Core Concepts Posts: 365 DataikerOptions
Hi,
Can you please send install.log (administration->maintenance->log files) for review? If you cannot share or do not want to share it, you can search for any traces of dssadmin command to find which ones were executed previously. This should give you understanding what was already in place with your existing DSS instance
Answers
-
Hi Sergey,
First of all, big thank you for the quick answer !
It is a EC2 instance on AWS, so I created a snapshot for both disks (separate root and DATA disk). Using the Dataiku Backup procedure is not really something I can do, because of the big size of the data folder.
Regarding the pre/post upgrade steps... since I'm only the sysadmin guy, I have no idea if the existing installation is using some or all of the features posted in the post-upgrade-tasks. Same about tasks-after-startup.
Is there any way to check if the current installation use any of those features (R, graphics exports, Hadoop & Spark, User Isolation Framework, containerized execution)?
Once again, thank you for the quick answer!