On Thu, 17 Dec 1998, David Harris wrote:
> It seems that you are saying that there are two states of the resync speed:
> (1) full resync speed which only happens after a full second of IO silence,
> and (2) the guaranteed minimum bandwidth resync speed. Are there any shades
> of gray in between? Any gradient speeds?
it will maintain _at least_ the speed of construction specified in
speed-limit. The minimum speed is maintained for the whole reconstruction
period, ie. for the time since resync started. So it will actually be
gradual if there are spikes in load and the reconstruction is 'in
advance'. The RAID code has no business to overrule the speed limit
though, so if the system is so busy that we have dropped under the limit,
it will maintain the guaranteed average speed.
also, you can change it dynamically, it's a sysctl after all ...
> Perhaps a background program could be written to monitor the scsi read-block
> and write-block statistics and properly adjust the raid resync speed-limit
> to keep access in some comfortable range which does not saturate the IO
> subsystem.
i dont think it will become a problem. The original resync was too
agressive and slowed down things like e2fsck. Probably it should not be a
'KB/sec' unit, but percentage of IO bandwith.
-- mingo
