Hello Theodore, Very interesting response, thank you! It all makes sense. And I realise it's hard to make everyone happy yes.
On 21-05-2020 16:48, Theodore Y. Ts'o wrote: > > In *general* things are mostly safe unless you need to do an > overlapping move of the inode table, in which case it is very hard to > make it be 100% safe. > Interesting too. And this may be in line with what I ended up experimenting with! Indeed I later realised I could just create LVM snapshots and e2fsck them: Always came out clean, while saying "filesystem with errors, check required". Of course I don't know whether data *integrity* was correct, but I presume that's "easy" to guarantee if everything's sequenced properly. That testing was all during the pass 2 block relocations. Which I guess is also the least risky pass then? > If you would like to help, you could try running resize2fs -p on a > test file system, and then try to randomly interrupt the resize at > various points, and then run e2fsck -fy and during which phase of the > resize you see things getting corrupted. > So I've tried the above 6×, and never managed to get it to report corruption, which is a great outcome as well! Pass 3 I didn't spot in time, and these were the only passes actually reported by the progress indicator. > It's not a high priority for me, since I have way too many other > things to worry about, if it is high priority for *you*, some > contributions to the effort would be much appreciated. After all, > Open source means you get to help fix the things you care about. :-) > Hah, for sure. It's not super important to me either and I think it's hard to do anything that *will* help some people while not being a backward incompatible change to others. Just like I presume telling people that interrupting is "probably" a safe thing to do during block relocation is risky for the next time someone does that and things *do* go wrong. However, with tools like this I always wonder why the progress indicator isn't on by default, and maybe in resize2fs' case a backward compatible change, and similar to dd, would be to enable the progress indicator on SIGUSR1? Kind regards, Wilmer v/d Gaast.
