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, Nov 25, 2004 at 07:05:33AM +0000, Cyril Bouthors wrote: > On 24 Nov 2004, Philipp Reisner wrote: > > Your SyncTarget node seems not to be able to handler this. > > Handle what? The throughput? yes. > > I would never configure DRBD to use more than 30% of my RAW disc > > trhoughput for synchronisation. -- This is just my oppinion. > > I completely agree with this opinion, I just wanted maximal throughput > for the initial synchronisation before going into production, now that > the servers are in production, I'm using a rate of 10M/s. > > With all due respect, I'm a bit more confident that there _might_ be a > bug somewhere in the synchronisation part because it has not been > widely used at full throttle. now, we used it, I have seen 70MB/sec, and it synced like a terrabyte or so. with one box being smp. and, we know that some combination of kernel/nic/nic driver just don't behave when stressed by DRBD. though we don't have a comprehensive list, not to speak of analysis of that. > > Since we got the AL high resync rates are not necessary at all... > > I don't get this, what's "AL high resync"? ... we have the Activity Log (AL), thus high resync rates ... lge