I have been running badblock's read-only test on the other bad drive (currently in the NAS) today. There have been a total of 292 bad sectors thus far - I don't trust this drive not to randomly fail/degrade rapidly.
If gsmartctl is right, the drive has been powered on for a total of around 7 years! Does this sound reasonable, Terry? Either way, we could try rescuing these drives with the destructive (or non-destructive) read-write options for badblocks, but I think it probably makes more sense to get a new 1 TB drive. I doubt they're too expensive, seeing as 1TB isn't huge these days. This is especially a good idea, because any custom programs/software we build/compile for this may have to be on the HDDs - the NAND storage is tiny and writing to it probably means re-flashing it. There is a serial/debug port on the PCB to which we can solder pins to unbrick it if we have to, but this may be an avenue best avoided. Also, the NAND is only 128MB - probably not enough to add anything useful. As for how well it handles damaged drives, the RAID re-sync failed (without any error!) after I removed and reinserted the bad drive. The current state of the array is "degraded" - we probably need a good drive/to fix one of these before we use it. Hamish On 15/06/2019 17:46, Ralph Corderoy wrote: > Hi Terry, > >> After logging these badblocks reported nothing until I gae up about 6 >> hours later. > `lsusb -t' shows the maximum speed the USB device may operate at. You > might improve things if you move it to a USB 3.0 port and it can cope. > >>> But this is a ‘new’ drive being added to the existing one to make an >>> array so it will be the destination of the mirroring and have lots >>> written to it? It's only the reading of it, e.g. a regular ‘scrub’ >>> that will show problems. >> The 'old' drive will have nothing on it, so we could put this one in >> first and then add the other. > But this one has bad blocks that can't be read. If the NAS does just > mirror all sectors from this to the second drive then it will have read > errors. Is there any data on this disk that you hope to save? If not, > write to all its sectors so the drive's firmware can abandon attempts to > read the dodgy sectors and start from scratch by then partitioning it > and making filesystems. > -- Next meeting: BEC, Bournemouth, Tuesday, 2019-07-02 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
