RAID patrol write to force sector remap/rewrite?

Asteroza

3/21/24, 1:16 AM

So I have a HPE S100i SR Gen10 software RAID 1 device that hosts 2 SATA SSD's for a Windows boot partion. The HPE Storage Array management software looks like it found a problem on the SSD's during a patrol read...

Warning: This logical drive has Unrecoverable Media Errors Detected on Drives during previous Rebuild or Background Surface Analysis (ARM) scan. Errors will be fixed automatically when the sector(s) are overwritten. Backup and Restore are recommended.

SSD's themselves are still fine apparently, only 25% write lifetime used. One has 0x0003e3dd read hard errors since factory, 0x00000e00 since reset, no write errors. The other has 0x00004c95 read hard errors since factory, no errors since reset, and no write errors. Strangely neither had read error retry or read error ECC corrected either. Same, no reallocated sectors.

At this point I think the patrol read hit some bad sectors that don't match on either SSD, so the read can "succeed" from the working side of the RAID 1 mirror, but I don't think it actually rewrote the bad sectors from the read. Which means if I tried to swap one side of the mirror with a new disk, the RAID 1 rebuild ends having an unrecoverable media error during rebuild and risks failing.

If the actual remap won't happen until a true write occurs on that sector, how can I force a write, to effectively flush the "pending" sector reallocations? Is there a windows utility that can force a full read AND write (read a sector, write the sector back) of the whole logical drive, similar to a ZFS scrub, rather than just a patrol read? Since The HPE RAID card is one of those software RAID things, and thus fake RAID, I'm not sure if non windows DOS utilities would work to force the write(back) (I think GRC SpinRite might be a candidate forced write utility, but apparently version 6.0 isn't safe for drives over 137GB due to a bug currently)

Since I haven't had write errors yet, there is the possibility that these are just empty sectors, and just running sdelete from sysinternals on the freespace would force the write, but there is no guarantee the problem sectors exist only in unwritten free space.

On the flip side, how tolerant of unrecoverable media errors are RAID rebuilds?

101

0 + 0

raid

software-raid

raid1

write

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: RAID patrol write to force sector remap/rewrite?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.