New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Comments
It sucks even for this. We're in the process of building a new shared hosting server because the current one on software raid is fine... except for disk i/o due to mdadm randomly verifying the array.
I guess the only viable use for mdadm RAID 1 is a desktop computer. It's going to cause you problems if you try to use it for any sort of server.
We use software raid 1 for our cPanel servers (not OAH), <300 accounts per server and using velociraptor drives. Been doing this for years, no problems at all even with software raid's random verify.
End of the day it's about how much spare disk IO and CPU the server has available during full load, since it's almost impossible to predict when the verify will run unlike for hardware raid where we can schedule it off-peak. I have a server (not OAH) running 80 VMs on 4x WD Black 500GB software RAID 10, very manageable because the VMs are mostly idle or very low usage.
Sorry to derail you guys but I also appreciate the responses you guys have given, kudos and thanks
RAID check/verify is done and everything's smooth again. How often does it do this check/verify process? Everything is fine on this server until it does that.
Thanks
Daily I think?
echo idle > /sys/block/md0/md/sync_action where md0 is the device should make it stop checking in the future.
http://ubuntuforums.org/showthread.php?t=1890796 has more info.
Er, I meant this link, but the other one I posted is good too: http://stackoverflow.com/questions/12114461/why-does-mdadm-keep-checking-resyncing
My /etc/cron.d/raid-check contains:
So I guess it's once a week.
Additionally, take a look at /etc/sysconfig/raid-check and read the top part of the file. It gives a bit of information on changing the 'nice' value to help with load issues.
Thanks,
I've implemented the command on md2 which is the big one that kept doing it... Hopefully this will be resolved
But you do realise that verifying the data is an important part of ensuring the integrity of the data right?
@shovenose If datashack is to expensive then thats not good at all, I would start rethinking your whole plan.
And software raid performance is still going to be 200MB/s and decrease once the node is fully loaded to 60-80MB/s.
I do have a couple spare hw raid controllers onsite, which will work fine with RAID 1, performance will also be consistant.