Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
> > Well, damn. As the program was supposedly reviewed by Samsung > engineers as part of their efforts to diagnose the root cause of TRIM errors, it > never occurred to me that it was that buggy. I can't thank you enough for > finding that! The rollout of some new DRBD clusters has been on hold for 2+ > months while we've been trying to track this down. > > Eric, thanks for the time you took to demonstrate and reproduce an issue > you thought you found. These things happen. Samsung engineers are people > too :-) although it seems they overlooked a rather elementary problem. > > Any many many thanks to Lars for taking the time to review this issue, finding > the cause and putting our minds to rest. > > kind regards, > Jan > I'm not quite decided. As Lars suggested, I'm going to try to make it fail a couple more times and see if the supposedly "corrupt" files are indeed >= 4GiB. I'll update everyone when I have more information. Thanks all of you for paying attention to this. --Eric