Nextcloudpi users on Odroid-HC1 and HC2 please read

ℹ️ Support 📦 Appliances (Docker, Snappy, VM, NCP, AIO)
1

I made a post here a couple weeks ago about an issue that I thought was serious enough to report as a bug, and I joined GitHub for the sole purpose of reporting it. The developer immediately closed the issue, responded with “you need a new hard drive,” and tagged it as “Use forums please this is not for support.” I emailed him to clarify the issue , because I thought he misunderstood me (and to let him know that the hard drive is actually NEW), but there has been no response. I’d like to find out whether other users of Nextcloudpi are seeing the same thing I’m seeing on the same hardware.

The issue I noticed is that, while the hard drive spins down as it should for power saving and to reduce wear, it VERY frequently spins back up, for no apparent reason, to the tune of 175-200 times per day. There is no setting available in Nextcloudpi that is adjustable by the owner/admin that would reduce the frequency of spinup. It is also not caused by a desktop sync client making a request, because I keep it disabled after the first sync of the day. It certainly seems that there is some underlying process, that is unavailable to the admin to adjust, that is frequently waking up an idle hard drive.

Start/stop cycles are an important limitation in mechanical hard drives, and a high start/stop cycle count can be a cause of early hard drive failure. An example is the very high failure rate of many so-called “green” hard drives a few years ago, due to frequent spindowns and spinups. The HDD in my device is only rated for 300,000 start/stop cycles, and at the present rate, that number would be exceeded it about 4 years.

I would like to find out whether others running Nextcloudpi on Odroid-HC1 or HC2 are experiencing the same high start/stop cycle count growth. To find out, you can run a SMART test under System/nc-hdd-test. Select Short Test, the apply. When it is finished (after 2 minutes), uncheck Short Test so nothing is selected, and hit Apply again to see test results. The count number with be the 4th item, with the number at the end of that line. Run the test again a day later and compare numbers.

If others are having the same problem, I’d really like you to make a report at the Nextcloudpi GitHub portal. Maybe that will wake up the developer to the problem, because it really felt like he was just blowing me off. I really like Nextcloudpi, but I’d rather not have an expensive hard drive fail because of it.

Thanks.

2

i guess it’s not only valid for HC1 and HC2… it’s just a systematic thing… caused by nextcloud (not nextcloudpi) which relies upon so-called cron-jobs. which should run at least every 5 minutes (as a default from nc 17+ on)

apart from that you could search the forum… i remember there were discussions about spinning down hdds, already.
if you’ll find a suitable oine pls don’t forget to post the link here that led you to the right path :slight_smile:

3

Thanks, @JimmyKater,

Is there any way for the user to adjust the frequency of these cron jobs? I’d like to be able to decide for myself how often the drive gets woken up.

It’s not an issue of a spindown timer - that is working well, and as frequently as I’d like. It’s how often the idle drive is being spun back up to satisfy these cron jobs, I guess, but I will search the forums for anything mentioning this.

Thanks again!

4

i’d suggest to read the linux manual about listing cron timers and adjusting them.

but be careful since you’re doing an operation at open heart of your nextcloud. if you don’t know for sure what you’d be doing i advise to leave it unchanged. it could harm your instance more than your hdd.

5

Thanks, @JimmyKater,

That may be a problem, however. Someone in my earlier post suggested changes I might make by logging in through SSH, but I was unable to do so. My login was rejected, and I was never even asked for credentials. It seems that any direct access to the underlying operating system has been closed. I can only access the settings that are available through the NextcloudPi web GUI, and those are quite limited. I have a feeling I’ll be at the mercy of the developer, but it doesn’t seem like he’s interested in looking into this.

6

Sounds like SSH may not be not enabled, you can enable it from ncp-web.

7

Now there’s a recommendation I hadn’t heard before. I’ll try that later today. Thanks!

8

Also is good to read the NextcloudPi documentation

1 Like