Connected HDD spins up constantly

[Lord Crisp]

Hey all,

Even if I'm a new user here, I've been lurking for a little while now... I just never had a reason to create a user, as I could usually find an answer to my questions.

I'm starting to go crazy from my external HDD connected to my RT-AX56U (Merlin 388.2_2) constantly spinning up...
Might sound like a minor issue, but my "extreme audiotory disturbance sensing" ADD is taking the life out of me.

I initially thought that maybe it was me having regular Check Disk setup, but looking at the logs, they're running once every 3 days or so...

My system log has a lot of messages from spinning up the HDD, and I can't seem to track what's causing it... here's a couple of examples:

sd-idle-2.6[1674]: spinning up /dev/sdb after 1 hours 9 mins 1 secs
sd-idle-2.6[1674]: spinning down /dev/sdb after 5 mins
sd-idle-2.6[1674]: spinning up /dev/sdb after 3 mins 1 secs
sd-idle-2.6[1674]: spinning down /dev/sdb after 5 mins 30 secs
sd-idle-2.6[1674]: spinning up /dev/sdb after 22 mins 31 secs
sd-idle-2.6[1674]: spinning down /dev/sdb after 5 mins

As you can see... the "spin up time" seems random...

I did see that there was some "talk" of ASD being an issue, but filtering by asd in htop shows me two processes, that I can't see doing anything "interestin"...
And from what I read in the thread about it, that was also about CPU usage, where mine is fairly okay (although I did have some issue a while ago, but seemed more like a malfunction somewhere, as a restart fixed it)....

Does anyone know what might be causing it? Or even an idea where to look?

I did actually look through the sd-idle-2.6.c on GitHub, but couldn't see what would "call it"... found where the message was created, and the the _manage function calling it (through more steps)...
Found it in the hope of maybe figuring out what the 1674 meant, but think I've just agreed with myself it's the ID that the router has given the drive, as there's another one for my USB stick that I have AMTM, Diversion, Swapfile and so on, installed on...

Looking through the log files, I've seen the "spinning up" message, all the way down to "after 31 sec".
And on average I'd say it spins up every 4-6 min...
And the longest in the log is after 4 hours and 48 min... but... long ones are rare.


My other external HDD, connected to my AC66U (used as a Mesh point in the living room) does not have this issue...

It's complaining about tntfs errors though, but... that not what this post is about

 

[ColinTaylor]

Are you sharing the contents of the USB drive over the LAN, e.g. via Samba? If so then devices on your network could be waking up the drive.

Do you have any addon scripts installed on the USB drive? They could also wake up the drive, if for example they write log files.

 

[DJones]

I would suggest a solid state drive if you have issues with sounds. Maybe try swapping hdd enclosures and see if it makes a difference. Some external exclosures including multi drive ones emulate the connections as iscsi and Linux systems can’t properly control them. Those types of external exclosures are better for short periods of runtime because they tend to have 100% spin up until you disconnect them.

Just a guess. Entware or a mapped smb drive can also cause that if your running programs from the drive and there is a lot of talk or logging happening.

 

[DJones]

Maybe unrelated I find logging from diversion causes a lot of slow downs and talk to the drive that contains the entware. Unless you need the logs they can be disabled and the ad blocker will still do its thing. Idk if anyone else has had this issue but the router in general has always run better for me with logging off. I use a cheap usb stick for entware.

 

[Lord Crisp]

Thank you guys for responding


@ColinTaylor
Yes, I'm running Samba on it.
But the only devices "firmly hooked up" to it, is the 3 computers of the house... and it's still spinning up when they're off.
And it would also seem a bit funny if that was the issue, because our other Samba (and Media Server) shared drive in the livingroom, does not spin up constantly.
Shouldn't this happen to both of them, if it was other devices waking it up?

Both our TV and Amplifier can see network drives, but I'm semi-sure it's only Media Server they can see... and that would then hit only the drive that is not a problem

And as I wrote in my post, I have a separate USB stick where I have all the "router programs" and stuff on, so all the scripts running and logs "written" happens on that one.


@DJones
The drives (HDDs) are both mounted as network drives on the 3 computers mentioned above.
But it's only for personal files, pictures and documents... we're not running anything off of the drive (nor the other one, which we actually use a lot)
Anything "running" is on the separate USB stick, with all the "router software" on it.

And... "Use an SSD" isn't helping my problem... even if I'd kind of agree, from a noise issue standpoint... this drive is specifically our personal things backup drive, that is maybe getting used once every 2 months or so... which is also why I thought it was okay to have it here "by my side", as it should never really be running...

And "spinning 100% of the time" also is not really a possible issue, when it's exactly spinning down, and then up again, constantly.

 

[DJones]

Tools like htop (under IO tab/disk read or disk write) or iotop or iostat can help see what activity and command is being issued.

I'm not sure if hdparm is compatible with asus routers, but it is a entware and can control drive power management spin up spin down time.

Alternatively disabling HDD hybernation will disable the spin up spin down and leave it always spinning. But again sound.

Looking back on the forums for older answers might also shed a clue on the problem of certain drives. Not sure if this is related but Rmerlin here seems to suggest drives internal firmware might have aggressive power management causing to spin up or down regardless of set disk power management settings on certain drives. https://www.snbforums.com/threads/spinning-down-hdd.19941/

iostat works also to see that exact disk.

(above pictures are from other devices not necessarily the router)

 

[Lord Crisp]

Tools like htop (under IO tab/disk read or disk write) or iotop or iostat can help see what activity and command is being issued.

I'm not sure if hdparm is compatible with asus routers, but it is a entware and can control drive power management spin up spin down time.

Alternatively disabling HDD hybernation will disable the spin up spin down and leave it always spinning. But again sound.
View attachment 51702

Looking back on the forums for older answers might also shed a clue on the problem of certain drives. Not sure if this is related but Rmerlin here seems to suggest drives internal firmware might have aggressive power management causing to spin up or down regardless of set disk power management settings on certain drives. https://www.snbforums.com/threads/spinning-down-hdd.19941/

View attachment 51703




View attachment 51701

View attachment 51704

iostat works also to see that exact disk.

View attachment 51705

(above pictures are from other devices not necessarily the router)

Thank you

I've never noticed the I/O tab in htop!
Already had it installed on the router, exactly for troubleshooting purposes


hdparm was installable with Entware, so I've done that now...
That thing has OPTIONS.
So I'll have to look into what to do and where, before I do anything crazy here...

Just wanted to update you on me seing your suggestion, so I wouldn't forget and seem ungrateful.

The iotop and iostat do not seem to exist for ASUS routers, through Entware...

And... to the rmerlin comment... that is actually interesting, and has likely narrowed the entire issue down to one thing...

Again, thank you!
Will try remembering to report back, when I've had time to look more at it It's 3:35 at night here right now


Edit:
Although...
Isn't all of this for the opposite problem... that people have problems with the HDDs idling when they don't want them to?
Mine is the opposite.

 

[DJones]

Thank you

I've never noticed the I/O tab in htop!
Already had it installed on the router, exactly for troubleshooting purposes


hdparm was installable with Entware, so I've done that now...
That thing has OPTIONS.
So I'll have to look into what to do and where, before I do anything crazy here...

Just wanted to update you on me seing your suggestion, so I wouldn't forget and seem ungrateful.

The iotop and iostat do not seem to exist for ASUS routers, through Entware...

And... to the rmerlin comment... that is actually interesting, and has likely narrowed the entire issue down to one thing...

Again, thank you!
Will try remembering to report back, when I've had time to look more at it It's 3:35 at night here right now


Edit:
Although...
Isn't all of this for the opposite problem... that people have problems with the HDDs idling when they don't want them to?
Mine is the opposite.

No worries

So hdparms idle time can adjust how soon the drive goes into idle after there has been disk activity. If the drive is a standard hard drive without any drive internal hybrid caching then any activity to the drive requires the drive to spin up either to read or write. If this was a Linux computer you could use preload to keep frequently used reads in ram but routers have little ram. One could probably delay aggressive writing to a drive by using system write through or write back caching however since the drive is being used on a router which has very little ram that’s not really an option. So the most you could do is either modify the time it takes to go back into idle or identify the cause of the activity. That’s the tricky part.

Unfortunately it really depends on what’s causing the activity because certain processes can’t be disabled without causing issues to file transfers. Someone suggested by using/installing asus’s download manager it caused increased disk activity you can chose to not install that entware and that might help, but if it’s Samba or Dlna media server they might need to be better configured or end devices that are connecting inspected if they are causing periodic reads to the disk.

If entwares install location or swap isn’t installed on this drive then logging and other processes shouldn’t use the disk besides samba, Time Machine, media server, or any third party entware you’ve installed like Transmission.

I’m not familiar with any kernel related activity that could be triggering disk activity. That’s not to say that nothing kernel or asus related couldn’t be the cause.

If it turns out to be samba their are a slew of global settings you can adjust under /jffs/smb.conf

smb.conf

Dead time might be of particular interest.

This global option sets the number of minutes that Samba will wait for an inactive client before closing its session with the Samba server. A client is considered inactive when it has no open files and there is no data being sent from it. The default value for this option is 0, which means that Samba never closes any connections no matter how long they have been inactive. You can override it as follows:
[global]
deadtime = 10
This tells Samba to terminate any inactive client sessions after 10 minutes. For most networks, setting this option as such will work because reconnections from the client are generally performed transparently to the user.

One other thing off the top of my head that might cause activity to a samba mapped drive is background scans or disk optimization from an antivirus program on a computer. Often they scan any connected drive when the computer is on and idle. How frequently this occurs depends on the antivirus software.

Other things could be something like Steam if for some reason you store your games remotely on the drive and there are game updates.


Things to consider about the sound of your drive.

The RPM of the drive and quality of sound/vibration dampening of the external enclosure can play a factor in how much sound is emitted from the drive. RPM between drives is not always the same faster RPM drives will result in more noise in the spin. Drive age, bad blocks, and fragmentation or many small files can also cause the arms to need to seek more which can cause clicking sounds within the drive.

A toaster based external drive enclosure for example might emit more sound from the drive then a full enclosure with rubber vibration padding.

 

[saleanddeki]

The same story with me, it started with version 388.x. This did not happen with previous versions. The same hardware, software and network configuration as before. Something in the router firmware does that

 

[SorenBalle]

Did anyone manage to resolve this issue?

I recently ditched my netgear nighthawk router and got an ASUS Rt-ax86u … and now have the same problem. The exact same 4TB 2,5” Tohshiba USB 3 hdd would power down/sleep/hibernate just fine on my Netgear router … it has now been running non-stop in my new ASUS Router for 4-5 days. (Yes I did activate hibernation).

I am afraid it will burn-out if it doesn’t power down now and then…

Any recommendations ?

 

[RMerlin]

I am afraid it will burn-out if it doesn’t power down now and then…

Hard drives are intended for 24/7 runtime. Spinning down/up multiple times per day will be far more damaging to the hardware. Just like a light bulb will almost always burn out when you turn it on.

And BTW... don't post the same question in three different locations. It will only cause a lot of redundant answers to be repeated in separate threads, and make any discussion impossible to follow.