update HRO clould configs after replacing the failing node #576
Conversation
crwood
left a comment
There was a problem hiding this comment.
Thanks for this.
A few comments based on previous observations:
"shares-happy": "3",
This means that shares must be placed on all three servers in order for uploads to succeed. Is this what you really want? Consider lowering this to "2" in order to allow uploads to succeed if one server is offline/unavailable (which might be wise considering this server has already been re-provisioned once).
"storage-server-FURL": "pb://35uweljc5o6tq2pbbwbbabvfpwqphicc@tcp:104.144.74.50:8898/5qspucvk47jd2653hlwuqjze6dgwqvvf",
Is there a reason why this fURL (and not the others) points to an IP address rather than a hostname? To make it easier to move if needed, maybe this one should use a hostname too. What do you think?
- Use hostnames for all fURLs instead of hardcoding IP addresses - Add/pre-allocate additional (offline) servers in the configuration (servers 4-5-6)
|
Tested and confirmed working locally -- thanks! |
update HRO clould configs after replacing the failing storage003 node