Reading data from raid with 'bad stripes'

[Lachran]

Not sure if it's proper forum for the following question, but will try it here first.

After dead disk was replaced in raid 5 and array was rebuilt, the controller repoted bad stripes on the array.
System (Linux) can mount and read all folders and some files from the array and fsck reports no errors during the boot.
There is one file in particular, which cannot be read. Any attampt to read returns with I/O error. File is rather big, about 60% of the entire array.
The question, is there a way to read at least some data out of that file ignoring bad blocks?

Any idea will be appreciated.

[Rubberman]

It probably means that another drive is starting to go bad since this sort of error on a RAID-5 array should only occur if some of the sectors on the old drives cannot be read properly. While fsck on boot may show "clean", you really need to do an fsck with the force option to make it check the drive even if it appears clean.

FWIW, it is not uncommon for more than one drive in a RAID-5 array to die within a short time of each other, since most such arrays are made with drives from the same manufacturing batch. This tends to skew the MTBF on the drives if any of them fail prematurely.

So, backup all the data from the array you can, then unmount it and run fsck with the force option. Then see if you can get the rest of the huge file that takes up so much space. What is it, BTW? 60% of a RAID seems a bit excessive to me. Was it a log file, or something that you really need to keep?

Finally, replace the rest of the drives on the array before you restore your data. Disc space is cheap (relatively) these days. A 7200rpm 1TB disc is about $100 USD.