Having below queries on how ES snapshot works
(ES Version: 6.2.1)
With a snapshot containing multiple indices, does the snapshot'ng process go in parallel across indices (looks so and expected), replicas and nodes. Tried different scenarios (all 2 shard indices)
a. 1 indice 20GB took 13 mts
b. 1 indice 40GB took 30 mts
c. 1 indice 90GB took 1 hr
d. 17 indices 120GB took 29 mts
If yes? how ES manages the I/O throughput for the storage. In our case, we use Azure Repository and it has its own scale by offering. Do we have any setting towards this?
When the snapshot was showing IN_PROGRESS against multiple indices scenario with API GET /_cat/snapshots/es_snapshots?v, the task API GET /_cat/tasks?v didn't mention any snapshot task. Though, it was showing when snapshot was done with other 3 scenarios. ALso tried, with GET _tasks
Thanks for clarification.