On a Windows system there is a unility called Crystal Disk Info -- it provides a first level analysis of the health of your hard drive. It may be a blunt instrument but it is better than nothing.
Is there some equivalent for the ReadyNAS 314 and 312 product?
Is there some equivalent for the ReadyNAS 314 and 312 product?
There are a several.
There is a built-in disk test that can be periodically scheduled or run on demand. That test uses the long SMART test that is built into modern disk drives. You will get a pass/fail indication in the main log, and you can look at the detailed SMART stats in the full system log zip file (downloaded from the admin web ui).
Also, the system does run the short SMART test on each drive daily. The results aren't reported explicitly in the web ui (unless a drive fails of course), but you can still see the detailed SMART stats in the logs.
In addition to the disk tests, you can also schedule defrag, btrfs balance, and volume scrubs from the admin UI. If you are using RAID redundancy, the scrub also exercises the disks (reading or writing every sector on the data volume), so it serves as a disk test.
One caveat - the ReadyNAS uses higher thresholds for reallocated and pending sectors than I am comfortable with. Unfortunately there is no way to set your own thresholds. So you do need to periodically look manually at the disk SMART stats.
Also, you should set up email alerts, so the ReadyNAS will notify you when it detects a problem.
If you go to System > Volumes and hover your mouse over a disk you'll see the ATA error count.
If you download the logs (System > Logs > Download logs) you can see current SMART stats in disk_info.log and you can see the history of changes in key values in smart_history.log
smart history log is definitely informative. i recently had a disk with reallocated sectors. i wish i knew about these details. definitely adding it to my to do's to check these twice a month atleast.
i run disk test, defrag, balance, scrub fortnightly. disk test takes the longest almost a day. is fortnightly a over kill and should i just run these once a month ?
Unless you have a tremendous amount of data churn on your NAS, those are way more frequent than necessary. I run a balance monthly, a scrub every four months, and a drive test never. I believe that the access performed by the other processes is sufficient to cause SMART to kick in if there is a drive error. The monthly balance may still be more often than necessary, but it doesn't tie up the NAS much and it runs in less time the more often you do it.
I run a balance monthly, a scrub every four months, and a drive test never. I believe that the access performed by the other processes is sufficient to cause SMART to kick in if there is a drive error. The monthly balance may still be more often than necessary, but it doesn't tie up the NAS much and it runs in less time the more often you do it.
The scrub should also access every sector in the volume, so I agree it also serves as a disk diagnostic.
Personally I run each of the maintenance functions (including the disk test) every three months (on a staggered schedule).
As far as the frequency needed for balancing, one clue is the amount of time the balance takes. If it takes a long time, then that is a clue that you should increase the frequency. Mine generally take less than 5 minutes - which suggests that I could lower the frequency.
I run a ... drive test never. I believe that the access performed by the other processes is sufficient to cause SMART to kick in if there is a drive error.
I should probably also mention that my NAS is huge and all those drive tests would take a lot of time -- the scrubs are already enough. My main chassis is a 12-bay rack mount that contains two RAID5 volumes moved from an RN516 and an EDA500 and I also have a 24-bay SAS expansion with another RAID5 from another eSATA expansion and a newer 12-drive RAID6. All total it's just under 100TB, so all the drives are 6TB or smaller. The every four months scrub plan is so each volume has it's own month, just to make it easier.
FWIW, though not germane to this discussion, I want to mention that the migration from the RN516 and two eSATA expansions to the new rack-mount monstrocity was seamless. I moved the drives and powered up, and all went without a hitch. Kudos to Netgear for it being so, especially since the SAS chassis is not from Netgear. Of course, I did try it out with smaller arrays with scratch drives and had a full backup before I went all in. I was very happy that a factory default and restore was unnecessary.
i have been running these monthly once. i now have changed to run disk test,scrub,defraf once in three months and i am running each of these tests on different months. my Balance test also takes 5 mins roughly and that i left it to month. does this sound reasonable?
@aniraghome wrote: i have been running these monthly once. i now have changed to run disk test,scrub,defraf once in three months and i am running each of these tests on different months. my Balance test also takes 5 mins roughly and that i left it to month. does this sound reasonable?
