Node UUID Should Have Matching Hostname

[Jervin R]

This is a feature request.

If you have ever looked at numerous cluster logs, it becoming ever difficult to follow what happens to a node through its UUID. We should at least append the actual hostname to any log entries that only exposes UUID to easily understand what is going on.

For example:

150620 15:24:56 [Note] WSREP: view(view_id(PRIM,27d17c11-de78-11e4-ac4b-9b1d3b53359c,140) memb {
 27d17c11-de78-11e4-ac4b-9b1d3b53359c,
 4537e123-175c-11e5-881c-ab2f57b2be23,
 77a4d77d-175c-11e5-b139-5aee302553f4,
 81b5a118-1760-11e5-99b3-7fc84e67eaca,
 ace88622-175c-11e5-a169-7e97e05e14ed,
 c84abd07-ff02-11e4-889a-d74f07929fce,
 e3e7dd2f-de7f-11e4-b8d2-96412c740023,
} joined {
} left {
} partitioned {
})

You can try to map these UUIDs to earlier entries in the logs, but if the node is cleared/bootstrapped and its UUID changes, we are back to square one.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXC-1185