Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2004-02-11 22:29:58 +0100 \ Andreas Huck: > Hi, > > On Friday 06 February 2004 04:40, Anshul Gupta wrote: > > Hi Everyone, > > > > I am in process of setting up HA cluster and using DRBD for filesystem > > replication. I have created partition /dev/hda2 of 20 GB, currently > > using just 8GB. It takes approx 30 minutes to sync the entire partiton > > for the first time. > > > > Can I change the disk-size parameter to say 10G (in drbd.conf) so that > > sync time is reduced and later when I need more space, increase it? > > Will there be any problem if I proceed this way? > > the disk-size in drbd.conf just needs to be larger or equal to the size > of your filesystem on top of it. So if you use reiserfs, for example, you > may do the following: > > 1) stop heartbeat on both nodes (/proc/drbd shows both as secondary) > 2) drbdsetup /dev/nb0 primary (on node A) > 2) resize_reiserfs -s 10G /dev/nb0 (on node A) > Beware: shrinking is experimental. If its a new system, just > create a 10G filesystem (and a filesystem which allows to enlarge it). > 3) drbdsetup /dev/nb0 secondary > 4) Comment out the original (or final) disk-size and set it to 10G in > your /etc/drbd.conf (hm, don't know whether it understands the G, > just copy-n-past $((10*1024*1024)) from your bash :-) uhm. please do 4) before 2), and make sure that the new device size shows up in drbdsetup /dev/nbX show ! yes, it understands G. > 5) scp your new drbd.conf to your node B. > 6) (re)start of drbd on either node will make it primary and start a > full sync.