nate wrote: > problems like this is why i believe 2.4 is not near > stable yet, and why i won't be usin it for at least 10-11 more > months on anything including test systems.
Your call, of course, for your machines. But in general I've found 2.4 to be pretty decent. I had something like two months of uptime with 2.4.9 before I decided to upgrade it to 2.4.12-ac3, which in turn ran for a few weeks flawlessly before I decided to upgrade to 2.4.15... which was a mistake, to judge from the postings we've seen today. Fortunately, I hadn't rebooted since starting 2.4.15, so the unmount bug hadn't bitten me. So I just went to single-user mode, synced, and rebooted into 2.4.12-ac3, forcing fsck just to be on the safe side (it didn't find any problems). In general, a policy of "wait a few days to see if any catastrophic bugs are found, then keep your last-known-good kernel around just in case" has performed very well with 2.4. Craig
