NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
SMART Extended Self Test. Why?
Noob question here: I understand that ReadyNAS does a short SMART self test every day and will report back if there are any issues in the SMART attributes. However, I have the ability to schedule routine extended self tests. I understand what the difference between the two is (i.e., that the extended test is a short test with a full disk surface scan), but I keep asking myself why that full surface scan is valuable to perform. The surface scan will find and mark any bad sectors all in one go, whereas the drive would usually only locate bad sectors when it attempts to access a bad sector. Either way, they end up getting remapped in due time.
Can someone please explain? Why should I run an extended self test at all? What is the advantage over just letting the drive do its thing and report in as it discovers issues?
If I should routinely run the extended self test, how do I determine how often this should be done? Most users report that they do them monthly on the forum, but is there a set of factors I should consider when creating my disk maintenance schedule?
Thanks for your help!
Learning2NAS wrote:
Can someone please explain? Why should I run an extended self test at all? What is the advantage over just letting the drive do its thing and report in as it discovers issues?
I have a lot of files on the NAS which are only rarely accessed, and my NAS backups are all incremental. So if sectors go bad on a section of the drive where those old archived files live, the NAS won't find those sectors for a very long time. So let's assume that I get some bad sectors like that on drive 2. Now drive 3 fails, and needs to be replaced. But when I replace it, the RAID array fails to resync, because the resync uncovers the earlier latent problems on drive 2. The result: I'd lose the entire data volume.
So that's why you want to run some test every now and then on drives in the array. You want to do what you can to discover drive issues quickly, so you minimize the risk of a double failure.
BTW, a scrub has a similar effect - it reads every sector in the volume (including parity blocks), so it also serves as a decent drive test. Defrag and balance don't read every sector (neither does a file system test). One difference - the SMART test covers the OS partition, and the scrub does not.
I run all the tests once/quarter. I don't see a need to do extended SMART tests and/or scrubs more often than that.
3 Replies
Replies have been turned off for this discussion
Learning2NAS wrote:
Can someone please explain? Why should I run an extended self test at all? What is the advantage over just letting the drive do its thing and report in as it discovers issues?
I have a lot of files on the NAS which are only rarely accessed, and my NAS backups are all incremental. So if sectors go bad on a section of the drive where those old archived files live, the NAS won't find those sectors for a very long time. So let's assume that I get some bad sectors like that on drive 2. Now drive 3 fails, and needs to be replaced. But when I replace it, the RAID array fails to resync, because the resync uncovers the earlier latent problems on drive 2. The result: I'd lose the entire data volume.
So that's why you want to run some test every now and then on drives in the array. You want to do what you can to discover drive issues quickly, so you minimize the risk of a double failure.
BTW, a scrub has a similar effect - it reads every sector in the volume (including parity blocks), so it also serves as a decent drive test. Defrag and balance don't read every sector (neither does a file system test). One difference - the SMART test covers the OS partition, and the scrub does not.
I run all the tests once/quarter. I don't see a need to do extended SMART tests and/or scrubs more often than that.
Hey Stephen,
Thanks for responding to my inquiries, again :)
Two more questions:
(1) You do the disk test and scrub once per quarter. Do you run them both together, or seperated by a month and a half?
(2) Does your idea of a maintenance schedule change as the disks appraoch the end of their lives? Say a disk was statistically near failure (7+ years old) or starting to develop bad blocks. Rather than pull it from the array, I would like to keep it until the end and just do whatever routine maintenance that requires. What is adequate under these circumstances?
Netgear has no official recommendations on this, but I schedule all the maintenance tasks once/quarter for each volume:
Disk Test: 1 Jan, 1 Apr, 1 Jul, 1 Oct
Balance: 8 Jan, 8 Apr, 8 Jul, 8 Oct
Scrub: 1 Feb, 1 May, 1 Aug, 1 Nov
Defrag: 1 Mar, 1 Jun, 1 Sep, 1 Dec
My OS6 NAS happen to have 2 volumes, the second one offsets the above schedule by 2 weeks.
Balance came along after I set this up btw, which is why it is squeezed in.
I don't adjust the schedule as the disks age. If a disk actually starts failing, I'll generally disable these tests until it is replaced.
FWIW, for the shares that use snapshots (which is most of them), I only go back three months. Every month I manually delete the oldest snapshot in each share. I'd love a way to automate that.
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
