This topic provides answers to some frequently asked questions about node freezing and unfreezing.
-
What is the difference between freezing a node and freezing an instance?
-
What happens when I freeze or unfreeze an auto triggered node?
-
What happens when I freeze or unfreeze an auto triggered instance?
-
What happens to data backfill and test instances after I freeze a node?
What is the difference between freezing a node and freezing an instance?
Freezing operates at two independent levels:
| Freeze a node | Freeze an instance | |
|---|---|---|
| Scope | Affects all future instances generated from the node, starting the next day | Affects only the specific instance |
| Effect on the other level | Does not retroactively affect already-generated instances | Does not change the freeze status of the node |
| Where to act | Operation Center or DataStudio | Operation Center |
This distinction is the key to diagnosing most freeze-related issues.
What happens when I freeze or unfreeze an auto triggered node?
Every night, DataWorks generates auto triggered instances scheduled to run the next day. When you freeze a node:
-
Instances scheduled to run on the same day are not affected — they run as expected.
-
Instances generated after the freeze are frozen and will not run. Their descendant instances are also blocked.
When you unfreeze a node, newly generated instances resume their normal schedule starting the next day.
Freeze a node in Operation Center
Freeze a node in DataStudio
In the Schedule section of the Properties tab, set Recurrence to Skip Execution to freeze, or Normal to unfreeze. Commit and deploy the node for the change to take effect in the production environment.
What happens when I freeze or unfreeze an auto triggered instance?
Freezing or unfreezing an instance does not affect the node it belongs to. If the node is frozen and you unfreeze one of its instances, the other instances scheduled for the next day remain frozen.
Freeze and unfreeze are manual operations. Check the operation logs if you need to review what was done.
What happens to data backfill and test instances after I freeze a node?
Data backfill node instances and test node instances are snapshots generated from an auto triggered node. Freezing the node also freezes any data backfill and test instances generated after the freeze. Instances generated before the freeze are not affected.
Why does a frozen node still run as scheduled?
This typically happens for one of two reasons:
The freeze only applies to future instances. The freeze operation does not affect instances already generated before the freeze. Instances scheduled for the same day the freeze takes effect run as expected. Only instances generated the following night onward are frozen.
Data backfill and test instances are snapshots. If you manually created a data backfill node or a test node before the freeze, the resulting instances are based on snapshot information taken at creation time. These instances are not affected by a later node freeze.
To confirm whether a node is frozen, check its status in Operation Center.
How do I rerun an unfrozen instance?
After unfreezing an instance, whether it runs depends on its scheduled time and the status of its ancestor node. Using Instance a (ancestor) and Instance b (unfrozen) as an example:
| Ancestor (Instance a) status | What happens to Instance b |
|---|---|
| Not yet run | Enters Waiting time state, then runs as scheduled |
| Currently running | Enters Waiting time state, then runs as scheduled |
| Successfully completed — scheduled time not yet reached | Enters Waiting time state |
| Successfully completed — scheduled time has elapsed | Fails to run. Click Rerun in the Actions column to run it manually. Descendant instances run after Instance b succeeds. |
| Failed | Not run. See Tasks not running to troubleshoot. |
How do I check who performed freeze or unfreeze operations?
View operation logs on the Cycle Task or Cycle Instance page in Operation Center.
