hi folks:
We are a VLDB shot with over 70TB using SQL FCI with 2 nodes plus a cloud witness. The pain is
around patching. Every time after patching, before reboot the node, we have to put into disk maintenance mode as per recommendation from MSFT. However it automatically triggered the repair job. For database with 70+ TB, we have to wait hours before we can simply reboot the node ?
Questions:
- Why can't we just skip this as only drain the node ? I've checked many places and there is no official document supporting but a few docs hinted that os level patching / sql server patching may not require disk maintenance mode.
- What's the downside if we simply go ahead rebooting the node while repairing job is in progress? Wouldn't this a scenario covered by HA which we pay the money for? Thanks Hui