Drives won't sleep or stay sleeping

[storageneeded]

Out of the box a brand new TS-253a without enabling anything, adding any apps, or even enabling a single share, won't let the drives sleep. This seems to be a common complaint but none of the apps or processes in this lengthy list are enabled:
https://www.qnap.com/en-us/how-to/faq/a ... andby-mode

Even with no LAN connection at all something is still accessing the drives randomly as often as every few seconds. If they do manage to spin down they just wake right back up again after less than a minute--again even with no LAN connection at all (nothing plugged into the NAS). This is with two brand new approved WD Red NAS drives in RAID 1.

This is my fourth Qnap NAS but the others are all running version 3.x QTS and all have let the drives spin down and stay that way. QTS 3.x typically has around 15 processes in the Resource Monitor and I was shocked to see over 80 with QTS 4.3 out of the box. Perhaps some are needed for the "OceanKTV - QNAP’s NAS-based karaoke system" proudly touted in the Qnap press release for the TS-x53a series? There seems to be a loss of focus on the core functionality of a NAS: Reliable storage. Something as basic as the drives being able to sleep when there's zero LAN access is a core feature of most any NAS or storage device. If Qnap has managed to break that I'll likely return the NAS as they don't seem to offer any firmware older than 4.2.x for it.

Does anyone have any ideas or suggestions?

[Ron1963]

So what did the diagnostic program told you? The one mentioned at the end of the link you included. There is a whole subforum dedicated to this topic, have you had a look there? viewforum.php?f=55

[storageneeded]

Thanks Ron1963. I haven't run the utility as it seemed like the results were only useful to Qnap. And, given this has been a widespread ongoing problem for quite a while over many releases, Qnap doesn't seem very interested in fixing it. Thanks for the link to the other forum. I'll so some more research there. Does your TS-253 spin down the drives properly? If so what firmware are you using?

[OneCD]

* Moved topic to "HDD Spin Down (HDD Standby)" *

[Ron1963]

Hi Storageneeded, I am on QTS 4.3.3.0378 Build 20171117 and all that multimedia sh*t (at least what can be toggled by the user) is turned off or removed. I have discovered some firmware versions before my 253 spins down when it has been rebooted and not yet used the power schedule to be turned on. I use power schedule to put my NAS to sleep (so not shutdown) at night and to turn it on in the morning. Then as of next day it's not working anymore. But if I schedule a shutdown instead of sleep it spins down again. I don't mind and have not further investigated as during the day it's working with LMS for my music almost all day and it saves only 12W per hour when spun down (I have intelligent power outlets which measure power usage at any moment and sent it to an app on my phone).

[storageman]

Give up on this.

There are so many features now keeping disks awake you almost render your box useless to get them to go into standby.

I have tried on TS-253A and can get it to work, but the list of apps/features to disable is enormous!

[storageneeded]

Storageman I think you've lost sight of the primary function of a NAS. It's supposed to be Network Attached Storage above all else. My Qnap NAS running QTS 3.7 has no issues with the drives sleeping. And, based on my unhappy results with my new Qnap TS-253A, I purchased a Synology NAS and, guess what, it has NO issues with the drives spinning down and staying in standby. This is CLEARLY a Qnap problem with later versions of QTS as my TS-253A won't let the drives spin down out of the box even with everything disabled but basic NAS services. This is a feature that Qnap promotes and is supposed to save energy, extend drive life, and reduce the risk of a drive failure, but it's a total fail as it doesn't work. It worked in QTS 3.7, and Synology makes it work today, but you really think the answer is to just give up? It seems to me the answer is to buy a Synology NAS that works as advertised. Qnap doesn't seem to understand the primary function of a NAS. The TS-x53A series is advertised as a small business NAS. But Qnap seems more interested in having dual microphone inputs for Karaoke than having the basic NAS functionality work right.

[dolbyman]

major point of drive wear is spindle start/stop(including thermal stress)...so disk wakeup and sleep is causeing premature disk failure rather than preventing it.

also i do not quite get how non sleeping drives are failing at preventing nas core functions ..that way your nas is available all the time ...and not only after a 10+sec spinup period

[storageman]

Storageman I think you've lost sight of the primary function of a NAS. It's supposed to be Network Attached Storage above all else. My Qnap NAS running QTS 3.7 has no issues with the drives sleeping. And, based on my unhappy results with my new Qnap TS-253A, I purchased a Synology NAS and, guess what, it has NO issues with the drives spinning down and staying in standby. This is CLEARLY a Qnap problem with later versions of QTS as my TS-253A won't let the drives spin down out of the box even with everything disabled but basic NAS services. This is a feature that Qnap promotes and is supposed to save energy, extend drive life, and reduce the risk of a drive failure, but it's a total fail as it doesn't work. It worked in QTS 3.7, and Synology makes it work today, but you really think the answer is to just give up? It seems to me the answer is to buy a Synology NAS that works as advertised. Qnap doesn't seem to understand the primary function of a NAS. The TS-x53A series is advertised as a small business NAS. But Qnap seems more interested in having dual microphone inputs for Karaoke than having the basic NAS functionality work right.

I don't disagree with you.

If you post the output of below we can probably eventually get your box to go standby. But as I said, you have to shutdown a lot of functions.
https://wiki.qnap.com/wiki/Find_out_whi ... m_spindown

[schumaku]

Yeah yeah yaeh ... can't wait for complaints that accessing the NAS data does take 8..15 seconds.

It's brilliant logic that no network cable connected should lead to a spin-down - not. Reading this all-over the net, so don't worry, you are not alone. The NAS will write logs, will try to send notifications, ... and some services will be retrying.

Provide the logs - even if you think these are good for QNAP only - so we can have an eye, please.

[storageneeded]

I think the difficult issue is there's no separate system drive for non-volatile storage (such as a small SSD) so the storage drives are used. The OS can use RAM drives for most system functions but then you would lose log files and such if the NAS is not cleanly shut down. I get that. But, as I said, it worked fine in QTS 3.7.

Qnap could address this several ways. If you enable HDD standby the OS could automatically disable services that will keep the drive awake, or it could grey out the standby option if you have services running that prevent sleep, or they could pop up a warning letting you know of what might keep your drives from sleeping properly. Any of those three options would be infinitely better than having it simply not work right out of the box with QTS 4.2 or 4.3.

For those who try to justify spin up/down is harder on drives than letting them spin 7/24 there's no evidence backing that up as long as it's not happening every 5 minutes. Plus the power savings is very real even if it amounts to only around $10/year for a 2 drive NAS. And the drive spin-up time, with a dual drive NAS and WD reds, is pretty reasonable at around 6 seconds.

In a home environment NAS access can be very infrequent. You might only need to access a NAS a few times a day to not at all if nobody is home. It makes much more sense, under those sorts of conditions, to let the drives sleep. I realize you can schedule the entire NAS to shut down at night but then startup is REALLY slow and requires using WOL if someone is up late and needs access. And that still only addresses about 6 hours/day out of 24 hours.

My point about basic storage is that is, and should be, the focus of a NAS. It's all the other stuff Qnap has added, 90+% of which most of us likely never use, that's causing the problem. It's like buying a sports car and it only comes with a camper on the back and snow plow on the front. I just want a car that works right and isn't hampered by a bunch of marketing-oriented extras that often don't work right anyway, create security vulnerabilities, and apparently prevent proper standby operation.

It's also very disappointing problems like this are simple to reproduce, widely acknowledged, yet remain ignored by Qnap release after release. Qnap also rarely or never responds to forum issues such as this one. Qnap sells $9000+ "enterprise" NAS units running the same buggy bloated OS and it's hard to believe any serious IT department would put up with it and Qnap's lack of support. One look at their security notifications page is all they'd need to buy something else.

Yeah I could probably get sleep working with enough work, hacking, etc. But the point is none of us should have to.

[storageneeded]

To revive this thread, I was asked for the result of the blkdevmonitor utility and here it is. If anyone knows how to stop md9_raid1 from constantly writing to the array every few seconds, and preventing the drives from entering standby, please let me know?

===== Welcome to use blkdevMonitor_v2 on Thu Feb 15 06:18:07 PST 2018 =====
Stop klogd.sh daemon... Done
Turn off/on VM block_dump & Clean dmesg
Countdown: 3 2 1
Start...
============= 0/100 test, Thu Feb 15 06:20:23 PST 2018 ===============
<7>[177318.861247] touch(12412): dirtied inode 2022 (volumeStatus.conf) on md9
<7>[177324.183455] kjournald(1971): WRITE block 541376 on md9 (8 sectors)
<7>[177324.183485] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177324.449722] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177324.464918] kjournald(1971): WRITE block 541384 on md9 (8 sectors)
<7>[177324.465098] kjournald(1971): WRITE block 541392 on md9 (8 sectors)
<7>[177324.679421] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177324.694032] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177338.902969] kworker/u8:2(10341): WRITE block 262608 on md9 (8 sectors)
<7>[177338.902998] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177338.934659] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177339.155951] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177339.174231] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177324.183485] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177324.449722] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177324.679421] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177324.694032] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177338.902998] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177338.934659] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)
<7>[177339.155951] md9_raid1(1960): WRITE block 1060216 on sda1 (1 sectors)
<7>[177339.174231] md9_raid1(1960): WRITE block 1060232 on sda1 (1 sectors)

============= 1/100 test, Thu Feb 15 06:23:54 PST 2018 ===============

[mtraherne]

I too am trying to get to the bottom of the qnap noisy drives. Also with a TS-253.

Here is an extract from my logs:

============= 3/100 test, Sun Nov 21 21:13:33 GMT 2021 ===============
<7>[51501.150715] kworker/u8:0(3173): WRITE block 809304 on md9 (24 sectors)
<7>[51501.150751] md9_raid1(2685): WRITE block 1060216 on sda1 (1 sectors)
<7>[51501.150768] md9_raid1(2685): WRITE block 1060216 on sdb1 (1 sectors)
<7>[51501.150751] md9_raid1(2685): WRITE block 1060216 on sda1 (1 sectors)
<7>[51501.150768] md9_raid1(2685): WRITE block 1060216 on sdb1 (1 sectors)

============= 4/100 test, Sun Nov 21 21:21:12 GMT 2021 ===============
<<<<<<<<7>[52458.013939] md9_raid1(2685): WRITE block 1060232 on sda1 (1 sectors)
<7>[52458.013944] md9_raid1(2685): WRITE block 1060232 on sdb1 (1 sectors)
<<<<<<<<7>[52458.013939] md9_raid1(2685): WRITE block 1060232 on sda1 (1 sectors)
<7>[52458.013944] md9_raid1(2685): WRITE block 1060232 on sdb1 (1 sectors)


Know what I find weird? Your logs and my logs are writing the same random block. Coincidence, fishy, or specifically a block used by qnap?

[mtraherne]

Know what I find weird? Your logs and my logs are writing the same random block. Coincidence, fishy, or specifically a block used by qnap?

Turns out that most of the logs for this problem show these two blocks. Can't be good for the drives to have the same block written every couple of minutes.

[testuserfoo]

253 user here too and its driving me nuts not spinning down. If i kill the crond service it does eventually sleep but changes to cron get lost after each reboot. Ive dug around the software and the architecture seems to hang off crond - seems odd way to do it. There is a ton of crap and as a software security bod i reckon there are a ton of vulns. Concur with posters above - its a poor product design and a working spindown/sleep is as tangible a requirement as being able to access files remotely over the network. ** poor.