Ephemeral storage for entire worker config & PQ?
We're using AWS's c5d instance types which include NVMe instance storage for CRIBL_VOLUME_DIR, but we're also paying for a small EBS volume to host the configs in /opt/cribl.
Is there a compelling reason to not just put CRIBL_HOME in the NVMe storage as well and have it pull down the config on startup?
I came to this question because I was thinking about a scenario where we might shut down instances during low traffic times and how to handle being sure the CRIBL_VOLUME_DIR on the Ephemeral storage was formatted and mounted. It made me think, "If CRIBL_HOME was also there, then I wouldn't have to worry about CRIBL starting up without CRIBL_VOLUME_DIR being formatted and mounted".
Answers
-
I should note that we're not using containers/pods at this point.
0