Scenarios Not Supported by AzCopy
While AzCopy is a versatile and powerful tool for copying data, there are certain scenarios where it may not function as expected or where its use is not recommended, these limitations can help customers avoid unexpected behavior and ensure smooth operation of their data transfer tasks.
AzCopy does not support running multiple instances simultaneously on the same virtual machine or physical machine. Attempting to execute multiple instances concurrently may lead to conflicts, resource contention, and potentially data corruption. For example, if two AzCopy instances attempt to write and to the same destination blob simultaneously, it can result in data inconsistency and corruption.
Let's delve into a scenario to grasp this concept more clearly. Lets say we're executing a copy operation between two containers while simultaneously, on another instance, conducting a similar copy operation between the different two containers. In this scenario, running multiple instances concurrently can lead to unforeseen and possibly problematic behaviors as can be noted below:
The First copy operations succeeds as can be observed
whereas the second copy operation gets failed
When attempting to copy data from a blob that is actively being written to by another agent, AzCopy may encounter unexpected behavior or fail to complete the transfer successfully.
To illustrate this scenario further, let's consider a situation where we're performing a sync operation between two containers. At the same time, on a separate instance, a removal operation is being executed on the destination container. This simultaneous modification of the destination container while AzCopy is attempting to synchronize data can lead to conflicts, inconsistencies, or even data loss.
The First sync operation succeeds:
Whereas, it can be clearly observed the second remove operation completes with errors resulting in inconsistencies:
