Hi Arnold,
Thanks for that input.
In your proposal, the old secondary will still be attempting to connect
to the primary as it isn't taken down. Is that likely to cause any
issues or will the primary just not accept the connection as its peer IP
doesn't match the expected one?
I like your general approach though. It begs another question:
If we snapshot a copy of the secondary over to the new node before your
step (5), will the sync process just copy the data between the time the
snapshot was taken and step (5) or would it still require a full re-sync
(hence having gained nothing)? I ask because the primary will have
successfully written more data to the old secondary than is consistent
on the new secondary - I'm wondering if that causes a problem confusing
the blocks of data to be synced from the primary (depending on the
internal structure of the metadata with which I'm unfamiliar).
I'd give this a try myself in a lab, but unfortunately don't have
sufficient hardware spare at the moment. Hoping someone with protocol /
"inner workings" experience will know the theoretical answers to save
some time.
Thanks,
Charles.
New e-mail signa
_______________________________________________
drbd-user mailing list
[email protected]
http://lists.linbit.com/mailman/listinfo/drbd-user
