Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Fri, Mar 19, 2004 at 10:02:17AM +0100, Philipp Reisner wrote: > May I ask you how you used drbdsetup/drbdadm to get it to crash on > this BUG_ON() ? I won't have access to the machine again until Monday, but from memory, I did: drbdadm attach all drbdadm connect all on both machines, then drbdadm primary all on one, and on that one, tried to make a filesystem. It hung the kernel almost immediately. Later on, I hit another BUG, also related to the extent parameters, but this time it happened right after I made it primary (I didn't have to make the filesystem) since it was trying to recover to the other machine. That crashed the sync thread, and /proc/drbd just showed it sitting at 0k/s. That BUG I could recover from by just shutting down the device. And ctrl-C'ing out of drbdadm while it was waiting to kill the already-dead sync thread. I will try to get it precisely replicated on Monday. I wonder if it's due to monolithic vs modular? I'm pretty sure I made the execution paths identical for both... -- Kees Cook Open Source Development Lab kees at osdl.org