Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2004-06-07 11:49:19 +0530 \ En Chiang Lee: > > Again, this might be problem, but I have to check it out. I do not know how > > much modification the SSI guys do to get the software integrated. I am using > > their tar file for this. But I will check it out this monday. > > >interessting. you may want to talk with the openssi guys, > > >they started to try and integrate drbd early this year. > > I did. They had some good suggestions, but it did not work out. This is the > > most frustrating part, most of the other guys who are trying this out, > > atleast the first sync goes through for them. For some reason, mine is not > > going through. > > Hi Arcot, > > We integrated drbd-0.7_pre5 with SSI. Currently, I've added a nodenum= > field in the drbd.conf to make it easier to make decisions during > startup since the hostname has not been set. > > However, we plan to move to drbd 0.7 as soon as it is formally released > and remove the dependency on node number and use the IP address instead > which is already part of drbd.conf. > > The changes in drbd are not much to get it to work with SSI, and we were > hoping to put it into the drbd cvs (Lars/Philipp?) once we remove any > user visible changes ( read nodenum= ). we are happy to even include youre nodenum, if properly ifdef'ed ... I'd already done that, only the drbd.tar on your download site is not even zipped, you use printk without severity prefix (what is legal, ok), you imported that tree into your own cvs, then modified parts of it, so it is difficult to see what exact version you based it on, you not even put a diff against the original source in there... and your small modifications are not even commented. you disable parts of the code, but don't say why. you put in an additional branch (e.g. in drbdd_init ) but don't say why, what the intention is, why it needs be done exactly there, or what "event" this reacts to... I'll mail you with private mail some diff where I think this might be your modifications forward ported to current cvs (or my working directory, for that matter, so tomorrows cvs). maybe you want to comment your modifications, and rather maintain a patchset than a fork. Thanks, Lars Ellenberg