Re: RAID 1 (mirroring) question

[Sven Hartge <sven@svenhartge.de>]

Bob Proulx <bob@proulx.com> wrote:

> The Linus software raid also had the capability to use a block bitmap
> to speed up resync after a crash because then it tracks which blocks
> are dirty.

> See the documentation on this mdadm command to configure an internal
> bitmap to speed up a re-sync after an event such as a power loss.

>  mdadm /dev/md1 --grow --bitmap=internal

Beware, this option can severly impact the write performance of your
RAID if your system is mostly doing small writes (which is what most
systems do).

My home desktop system was configured with a RAID1 with write intent
bitmaps and the subjective day-to-day performance was awful, like
driving a car with active handbrake. After disabling the bitmaps the
write latency went down and the system felt much more "snappy" and
responsive.

I now use an intent bitmap only an huge RAID sets which are mostly
written in big chunks (for example backup files from Bacula) and would
need a resync in the magnitude of days if it ever happend to get
desynced due to a power failure in the wrong moment.

Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.