Re: Several bugs/flaws in the current(?) bcache implementation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Nov 11, 2019 at 9:00 PM Coly Li <colyli@xxxxxxx> wrote:
> It seems Areca RAID adapter can not handle high random I/O pressure
> properly and even worse then regular write I/O rate (am I right ?).

I believe what his data shows is that when the only workload is
writeback, it's all random write and isn't coalesced.  Areca is pretty
good with random write, but because we sort on LBA already the adapter
cache provides no benefit to allow further I/O coalescing.

OTOH, I liked the system we had before where we had at most 1 idle
writeback outstanding at a time, because it was guaranteed to not
create excessive load on the cache or backing device.  I think his
chief objection is that he uses the cache device for other things and
doesn't want its idle utilization to be high.

Mike



[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Linux ARM Kernel]     [Linux Filesystem Development]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux