I like the solution of remounting. I'm running an extended SMART test on the drives, but after that I'll give it a shot.
The other option that gets my attention is "stripe=96", but I'm not sure I understand what that does in a RAID 1 array. Is this another option I should change? Hamish On 27/08/2020 13:15, Ralph Corderoy wrote: > Hello Hamish, > >> * The commit interval is set to 30 seconds instead of the default 5. >> * The data mode is set to writeback >> * barrier is set to 0, which, from what I understand, can also >> compromise data consistency. > The NAS is going for maximum performance and never mind the integrity of > the data. Could you remind us of the manufacturer. :-) > >> My understanding of these issues from what I read here >> https://www.kernel.org/doc/html/latest/admin-guide/ext4.html#options >> is that this effectively disables journaling, and seriously impacts >> the reliability of the storage system (RAID 1). > The journal will still be used, but in a way which makes recovery less > robust. > > ext4(5) also describes the options and the defaults of > ‘data=ordered,barrier=1,commit=5’ are much better. Certainly normal for > a database's filesystem, or anything where data loss is to be minimised. > > And RAID 1 doesn't help because these options are all about the > higher-level meaning of the data sent to the mirroring drives so > assuming the RAID 1 does it's job then you've multiple reliable copies > of incorrect data. :-) > >> Unfortunately, it is D-Link's scripts that mount the RAID array, so >> I doubt that I can change the options. > Look at ‘remount’ in mount(8). Not as part of binding, but standalone. > With the right arguments you should be able to change the mount options > to more sane ones soon after boot, with it ignoring /etc/fstab's > settings. >
signature.asc
Description: OpenPGP digital signature
-- Next meeting: Online, Jitsi, Tuesday, 2020-09-01 20:00 Check to whom you are replying Meetings, mailing list, IRC, ... http://dorset.lug.org.uk New thread, don't hijack: mailto:dorset@mailman.lug.org.uk
