Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Thu, Sep 10, 2015 at 02:21:02PM +0200, Joerg May wrote: > Hi, > > I have some questions regarding the initial syncing of a drbd partition, > where I could not find answers for in the drbd documentation. > > Let's assume the following case: > Two computers have their own hard disks, each of them has a > partition which is synced via drbd > with the other computer's hard disk partition. > Let's say this partition is only used to one third, so two thirds of > it are still free. > Now on one machine the hard disk has a physical problem and is > replaced by a fresh one. > Of course now for the fresh hard disk an initial syncing needs to be done. > > Now my questions: > - How efficient is the initial syncing ? > Will only the used blocks be synced or the complete partition (so > also the unused, free blocks) ? DRBD is agnostic to "usage", so it has no knowledge of "free". > Would it be useful to delete unnecessary files before the initial > syncing to reduce the syncing time ? No. > - Are there any means (e.g. drbd configuration options) to shorten > the initial syncing time ? Not really. > Would e.g. checksum-based syncing help to identify faster the free > blocks and skip them for syncing ? No. You may consider "truck based replication" (see the User's Guide). Or you can (if your backend supports "discards" resp. "trim") run an "fstrim -v" against the DRBD mount point on the SyncSource while the resync is running, that should usually speed it up (because it won't read data -> transfer data -> write data, but simply send "discard this LBAs" and then do so...) -- : Lars Ellenberg : http://www.LINBIT.com | Your Way to High Availability : DRBD, Linux-HA and Pacemaker support and consulting DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed