Reduced accelerator optimization in Azure for OR query with multiple tags
Explanation
When the Cloud object store policy has at least one query with multiple tag conditions, combined using the "OR" operator, the backups of Azure containers using accelerator enabled Cloud object store policy, it shows a loss of acceleration or backs up unchanged data.
This happens as the ordering of objects across multiple tags is not as expected for accelerator. Few objects that are not found in the tracklog even if they exist in the tracklog; hence, such objects are backed up repeatedly without getting accelerator benefit for these objects.
Workaround
Do not use the OR condition while combining multiple tag conditions for Azure. Instead, create a separate query per tag.
For example,
Say that you have the following query: (tagKey eq 'type' and tagValue eq 'text') or (tagKey eq 'type' and tagValue eq 'none) with say queryname datatype
You can create two queries say by name datatype-text with query (tagKey eq 'type' and tagValue eq 'text') and datatype-none with query (tagKey eq 'type' and tagValue eq 'none')
Note:
This results in the first backup, which is without any acceleration for these new queries. For subsequent backups, you can see the problem is resolved.