Volume is set to run a balance each month; probably the last time a balance was run would have been the beginning of december, at which time I was running OS 6.2.x.
Since those logs, I deleted ALL snapshots, the volume has 1.92 TB capacity and 816 GB free (stats from Frontview).
In order to keep things moving along I have manually stopped the balance a number of times to allow the nightly backup to run, etc, and then each day I'm restarting it. (been going around 3 days so far)
It looks a lot like the balance on 6.2.x was not running, or was different, so that this time around balance is a major hit. I'm hoping it will all finish in about another 8 - 12 hours, after that I'll run it again but based on what I see so far it looks like it will be much faster - once its got through this initial hump.
This initial poor performance is unfortuate though. Would certainly be desirable if it could either run faster, or run with less of a performance hit. This is a business critical machine for me, I'm curtailing my PC -> NAS backups during this time, which is not something I'm very comfortable with.
My advice would be to schedule defrag, balance etc. to run at a time when you have an extended period that the NAS would be used lightly, like you would with scheduling backups.
If run regularly defrags and balances should be pretty quick.
I did a defrag yesterday (stopped the balance to do it); this took only about 10 mins.
I have the balance to run monthly and it has been that way since OS 6.2; but I think there must be some kind of change from 6.2 -> 6.4, which means that now the balance has about 18 months to catch up on.
I'll know more about performance in another few days once this dust has settled.
A month after the hassles with balance taking ages, the next monthly cycle of a scheduled balance has begun.
The balance started at 11:00 pm. Its not 9:00 AM the following day (so balance been running for 10 hours), Frontview is barely responsive (mostly not); the balance is 29 out of 174 chunks balanced (274 considered, 83% left).
This NAS is business critical, it runs a subversion system and critical backups all day long (night times its not so busy).
This performance is TERRIBLE!
Please, can this be elevated on the priority lists.
I'll now have to try and do a balance shuffle as I did before (a little at a time, manually, using the command line) until it is done and then try changing to a scheduled weekly balance.
But really, this is unacceptable. Please please please look into it. Back in OS6.2 this was never a problem. Something has changed.
Ok. Some of the maitenance checks don't prevent spindown (though I'd be surprised if balance is one of them).
Running a daily balance seems excessive to me. I'm not sure what Netgear has set the dusage parameter to, but your long running times might be because it is set a bit too high.
I was doing a Monthly balance, and this was running for 2 - 3 days straight (starting on a weekend). This was still running on the Monday or Tuesday following, making the machine unusable.
I dropped the frequency to weekly, and this cut the duration of running down to around 24-36 hours.
I also found just after a balance, that running another would take about 10-15 minutes, so the idea of the daily balance was that this should be shorter, maybe an hour or 2 (I had hoped) - thus avoiding the long running times of the weekly / monthly.
Alas it was not to be. The daily balance has ended up with running times of 12+ hours, again making the machine unusable. I have a window from around midnight -> 7 am where things like balance, backup, scrub, etc can run. Apart from the the machine needs to be available and responsive.
So at present I'll give up on balance completely. It's not 100% required unless I find myself running out of space, so I"ll live without it.
As I pointed out before, balance was run monthly back on OS 6.2.0 and it was fast - I did not have any issues like this for nearly 12 months. Something changed between those OS releases, and now balance performance is dreadful.
