Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Nate Carlson wrote: > > On Wed, 6 Apr 2005, Todd Denniston wrote: <SNIP> > > 2) did you know that when you setup drbd against the disks the first > > time it has to do a full sync between the drives to get them to the same > > state, i.e. all 424334M has to be transferred? This could mean your > > mkfs.ext3 is causing the disk to seek between where mkfs is writing and > > where drbd is reading (these might be quite a bit apart) and this slows > > both down. You might try putting the mkfs to sleep and see what the sync > > rate goes to. > > I also tried sync'ing without running the mkfs.ext3 on them; it was > getting better speeds (~15MB/sec), but still not what I'd expect. > > I think I've got another problem with the drive array config on both > systems; I'm working on trying to figure that out now. Just remembered a trick that made my Gig ethernet cards a bit better, up the MTU on the cards (bonded pairs in your case) on both machines. For me `ifconfig eth1 mtu 6000` works to get the sync speed up from ~15MB/Sec to ~25-30MB/sec. There have been discussions on this list of how to optimize the MTU but I don't feel like searching for it right now. -- Todd Denniston Crane Division, Naval Surface Warfare Center (NSWC Crane) Harnessing the Power of Technology for the Warfighter