Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2006-05-10 09:50:35 -0700 \ Martin Fick: > > what, if you had an existing single-server setup > > with existing data, and you extend it with drbd to > > form a cluster, so you have existing filesystems > > with important data on it? is it ok to skip the > > initial sync there, too? > > Umm, no? Drbd would know that though right, since it > would find existing metadata on the single server > partition? no. never been drbd there before. never been drbd meta data. drbd8 drbdmeta recognizes some common file system formats, and requires confirmation if it thinks you do something stupid, but thats not bullet proof, either. and it is likely that someone will have a setup where there appear to have been a file system on both nodes before, when initializing drbd for the first time. this all boils down to: something may appear the "Right Thing" to do in the setup you consider to be the "standard way thing are", but the guy next door has a completely different view about what that "standard" typically looks like. > ... > > but then again, once we implement the online > > verification, we would _need_ to have this initial > > sync, because the drives _have_ to be bitwise > > identical > > (thats one of our "guarantees", after all). > > Ah, but that's where my intial zeroing suggestion > would help. While it might take some time to zero a > whole paritition (on both sides), there would be no > network tranmissions required... vastly faster than > syncing. well. no. only faster when you have a slow link. with a fast link, I typically see the drbd initial full sync finish slightly faster than dd would. > Please don't take any of this the wrong way, this is > by no means something that I *need*, and I certainly > do not want to suggest comlicated stuff. I think that > this would be rather simple and safe and I figured I'd > suggest it to see if it others thougth that it really > would be safe and simple, of course not, always glad to have some _usability_ feedback, I know usability of drbd sucks in many ways sometimes (I usually find out about this if I do a "workshop/handson drbd install" at some customers site...) any "it works but I think it could even work better this way" brightens up the usual routine of "this does not work, my drbd setup is broken, what did I do wrong" mails :) It just sometimes takes some time to convince me that a certain feature does indeed improve usability, while still being a true generic solution. cheers, -- : Lars Ellenberg Tel +43-1-8178292-0 : : LINBIT Information Technologies GmbH Fax +43-1-8178292-82 : : Schoenbrunner Str. 244, A-1120 Vienna/Europe http://www.linbit.com : __ please use the "List-Reply" function of your email client.