0

I run a small LAN and use hMailServer to provide an IMAP server. I have hMail's database out on a NAS drive. When I have hMail set to fetch email from external servers (eg. our ISP), it wakes the NAS disk. Actually, it never lets the drive go to sleep since I set it to fetch every 5 minutes (which is an eternity when you're waiting for email that you know is sent). My concern is over excess wear on the drive since it will never get to sleep (currently, it only sleeps if inactive for 60 mins).

What I'd really like is a method of allowing the client to trigger the retrieval of email from external servers and not have the server do this on any set interval. Is there any such mechanism that I could investigate?

Additionally, how much excess wear are we talking about with the server accessing the disk every 5 mins vs. every couple of hours? Is the disk still being worn if it is not being accessed, but also not in sleep mode?

2 Answers 2

1

It really depends on the hard drive and the duty cycle.

Lets start with the gorilla in the room. Hard drives do fail - and gradual wear is the least of your worries since you might still have time to back up everything and restore it. That said, enterprise drives are designed for 24 hour duty cycles, and lots of companies use large arrays of consumer drives for storage. Google has a paper on this, and the part of interest to you is

"When we look at these results across individual models we again see a complex pattern, with varying paterns of failure behavior across the three utilization levels. Taken as a whole, our data indicate a much weaker correlation between utilization levels and failures than previous work has suggested"

(section 3.2).

In short, drive wear dosen't really matter as much in terms of reliability as you'd think, especially where you may have other failure modes. Considering you're using a nas, it should be monitoring for wear related failure modes (using smart) and have some mechanism for replacing the drives.

I've tended to see 3-5 year lifespans on consumer hard drives (and they run 24/7 with a lot of disk activity, so there's periods where they're constantly reading or writing, or not doing anything), and when they don't have major electromechanical failures (which are often unexpected), they just slow down gradually. Keep an eye on your drives (smart tests are nice), have backups (all hardware dies) and you should be fine.

0

You'll have far more "excess wear" on the drive by starting & stopping it all the time, than by leaving it running, but it will consume more energy. The best way to wear out most electronics items is by turning them on & off. I suggest checking your hard drive model to see what its MTBF is; servers should use drives that have MTBF in the range of 10 years.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .