On 10 Mar 2019, Wols Lists uttered the following: > I'd like to modify the raid layer such that it times out quickly, and > recalculates and rewrites the data after a few seconds, such that these > drives cease to be a problem, but stick that on the long list of raid > papercuts I'd like to sort out when I can find the time to learn to > program the raid subsystem! I don't see how that could work. When these drives get stuck on lengthy retries, they are essentially unresponsive: often you can't even tell them to abort the command (and on some old drives they even locked the PCI bus up while they did it, though I haven't seen *that* for a very long time and it's probably impossible on PCIe anyway). If the drive is unresponsive while they retry, that means you can't ask it to write other data instead: it's unresponsive! You'd need to schedule a rewrite for later, somehow (not using the usual block-layer queueing, because if the drive times out and gets reset I'm not sure what happens to the contents of those queues). -- NULL && (void)
