LCC issue "automatically solved" by DSM or ... ?

[maxxfi]

I don't have an external enclosure so I cannot test it myself, but the synodiskport command help reports that you can use it with option "--sata" for internal sata devices, and with "--esata" for external sata devices.
So I suggest that you try the command:

Code: Select all

/usr/syno/bin/synodiskport -esata

If it seems to report the disks in the DX510, then to check the timers you can use:

Code: Select all

for d in `/usr/syno/bin/synodiskport -esata` ; do echo "*** /dev/$d ***"; /usr/syno/bin/syno_disk_ctl --wd-idle -g /dev/$d; done

And at this point if this returns some non-zero for the WDidle timers, use the following to disable them:

Code: Select all

for d in `/usr/syno/bin/synodiskport -esata` ; do echo "*** /dev/$d ***"; /usr/syno/bin/syno_disk_ctl --wd-idle -d /dev/$d; done

[JohnD76]

Hi Maxxfi,

This seems to enumerate the two esata ports but not the DX510's sadly. I will email Synology support and see if there is anything else I can try, but it appears that the script may not take into account expansion bays at present.

Thanks once again Maxxfi.

John.

[maxxfi]

Hmm... ok. One more idea. As Synology says that all disks contain a copy of the DSM,
and it does this my creating a big RAID 1 mirror, let's see which disks are member of that array

So, if you type the command "cat /proc/mdstat", check the part that begins with md0:
As example, this is the result with my NAS:

Code: Select all

md0 : active raid1 sda1[0] sdc1[1]
      2490176 blocks [4/2] [UU__]

In my case the two disks are sda and sdc.
See if you can find some other numbering that you didn't find with the previous tests.

[phils]

Because JohnD76 didn't reply.. I'll show you the mdstat for me 1511+ with a DX510:

md3 : active raid6 sdb6[6] sdgc6[7] sdgb6[5] sdga6[4] sde6[3] sdd6[2] sdc6[1]
4883713920 blocks super 1.2 level 6, 64k chunk, algorithm 2 [7/7] [UUUUUUU]

md2 : active raid6 sda5[0] sdgc5[8] sdgb5[6] sdga5[5] sde5[4] sdd5[3] sdc5[2] sdb5[7]
5832165888 blocks super 1.2 level 6, 64k chunk, algorithm 2 [8/8] [UUUUUUUU]

md1 : active raid1 sda2[0] sdb2[1] sdc2[2] sdd2[4] sde2[3]
2097088 blocks [5/5] [UUUUU]

md0 : active raid1 sda1[0] sdb1[1] sdc1[4] sdd1[3] sde1[2]
2490176 blocks [5/5] [UUUUU]

So it looks like the DX510 drives (for me) are sdg*

But they don't show up via synodiskport at all.

[phils]

I tried to run your command manually like this:

/usr/syno/bin/syno_disk_ctl --wd-idle -g /dev/sdga

But all of the drives returned:

Get /dev/sdga WD idle3 timer [138] successfully

I manually disabled wdidle (via dos bootdisk) on all of the drives before putting them in the 1511+ and the LLC according to smart looks fine.

[maxxfi]

But all of the drives returned:

Get /dev/sdga WD idle3 timer [138] successfully

So the timer is set to 138 (seconds? not sure what's the unit)
Then, use the command with the -d option "/usr/syno/bin/syno_disk_ctl --wd-idle -d /dev/sdga" to disable the timer, then reboot and check again with the "-g" option that the timer stays at 0.

[phils]

Ahh I think I worked it out. When I "disabled" the timer on my PC before putting the drives in my NAS I didn't disable the timer.. I set it to 5 minutes (which achieves the same result). That's why it shows 138 for the ones in the DX510.

What this tells us is that it isn't disabling the LLC on drives in the expansion units.

Disabling it via the command you provided works and survives a reboot. So it's sort of a bug I guess.

[PascalZ]

Speaking of bugs, I think I put my finger on a new one thanks to this thread.

I have a DS212+ running DSM 4.0-2219
Inside, two Western Digital Caviar Green: WD20EARX
I do NOT want the disks to hibernate so I set hibernation to OFF

The problem is that when the system is idle, the disks click every 10 to 20 seconds (and the disk led flashes). I understand that this is linked to the Advanced Power Management (APM) that is built into the disks. This problem was apparently fixed but it seems to be back, at least in my case (is it because of the DS212+, of the WD20EARX or the DSM 4.0-2219 ????).

Using the info in this thread, I think I have been able to determine that the WDIDLE3 implemented by Synology in the DSM is NOT working in my case (DS212+ with 2xWD20EARX under DSM 4.0-2219). I believe this is the root cause of the problem and it is probably impacting a lot of people worldwide.

For example, when I ssh to the Diskstation and I type:

Diskstation> /usr/syno/bin/syno_disk_ctl --wd-idle -s80 /dev/sda (or sdb)

i get:

Set /dev/sda WD idle3 timer [80] successfully

but when I power-cycle the system and ssh again and type:

Diskstation> /usr/syno/bin/syno_disk_ctl --wd-idle -g /dev/sda

i get:

Get /dev/sda WD idle3 timer [0] successfully instead of Get /dev/sda WD idle3 timer [80] successfully

The Diskstation can't set the idle3 parameter and this explains the clicks.

Have someone else noticed it too ?
Anyone, any help ?

Everything in my setup is standard "out of the box". And the WD20EARX id officialy compatible so I really don't understand.

Any help is welcome,

Pascal

[Lex]

Hi there,

I also have a new issue with LCC regarding my DS411slim (4.0-2219) and a WD WD5000BEKT-22KA9:

there are 4 drives present: 3x WD5000BEKT-00KA9 and 1x WD5000BEKT-22KA9 in RAID5;

after installing the drives in the DS the idle-timer was automatically set to "disabled" by the DS (you can verify that by manually setting the timer to any other value with a PC and and boot-disk with WDIDLE3; after placing the disks back in the DS the timer will be disabled again.)

This works well for the slightly older WD5000BEKT-00KA9; but with the slightly newer WD5000BEKT-22KA9 it does not work as it should: this drive won´t turn off the timer even it reports it correctly: when I read out the status with WDIDLE3 it says "disabled", but the drive somehow ignors this setting and keeps parking the head constantly every few minutes, the LCC increases fast and there is also some short "chirp"-sound every few minutes.

Now, my idea is, that this "22KA9"-version of the drive can´t handle the input of the WDIDLE-utility correctly, at least the disable-command. I want to check wether I can set the timer to a longer value (eg. 300 seconds) and this would fix the LCC-issue.
BUT: when I set the timer to 300 with my PC and boot-disk and install the disk back in my DS, the new value will be overwritten again by the DS and set to "disabled"...

Does anyone has an idea how to stop the DS from setting the idle-timer on the disks? Can I temporarily disable this command in my DS or modify the command to set the timer to 300 instead of turning it off completely? And how could it be done? I am not this familiar with linux; I can login and execute commands, but I would need a little walkthrough.. Thanks!

Alex

[PascalZ]

Hello,

Yes you can !
There is a script that is called each time the system is shut down that changes the IDLE3 parameter to "Disabled" on all the WD drives. The script is part of the DSM and can be found at: /usr/syno/etc/rc.d/S99WDidle3Dis.sh

If you edit this script and change /usr/syno/bin/syno_disk_ctl --wd-idle -d to /usr/syno/bin/syno_disk_ctl --wd-idle -s300 it will set the timer value to 300 instead of disabling it.

Don't forget to shut-down your DSM after that before turning it back on to do the tests.

I did it on mine and it didn't change anything. I hope you'll have more success. I'd be interested to know if it works for you.

Also, if you want to check the value of the IDLE3 parameter directly from the Synology, you can do it from the terminal: /usr/syno/bin/syno_disk_ctl --wd-idle -g /dev/sda (and sdb and sdc etc...)

Hope this helps,

Pascal

[maxxfi]

This works well for the slightly older WD5000BEKT-00KA9; but with the slightly newer WD5000BEKT-22KA9 it does not work as it should

Look that in the Synology HDD compatibility list there is a note for the WD5000BEKT:
"The batch number, 00KA9T0, for this particular hard drive has passed the hard drive compatibility test."

[Lex]

Pascal, thanks for the help - I will try this this evening, when I am back home.

I also found a completly different approach to stop the drives from parking their heads:

Disable Advanced power management using standard ATA command (Uses more power as turns off all low power modes but results in no load/unload cycles):
Linux users add following (hdparm -B 255 /dev/sdX where X is your hard drive device).

Here is a link how it can be implemented in the NAS so the settings are activated on every boot of the NAS via autostart: http://www.synology-forum.de/archive/in ... 0b19d908b7
(unfortunately it is in German language but I am sure you can understand how to do it)

I am not quite sure but I think hdparm is already present on the Synology NAS by default; if not, I read it can be installed.

Perhaps this will work if for any reason the WDIDLE-utility won´t do the job corectly?

Alex

[PascalZ]

Pascal, thanks for the help - I will try this this evening, when I am back home.

I also found a completly different approach to stop the drives from parking their heads:

Disable Advanced power management using standard ATA command (Uses more power as turns off all low power modes but results in no load/unload cycles):
Linux users add following (hdparm -B 255 /dev/sdX where X is your hard drive device).

Here is a link how it can be implemented in the NAS so the settings are activated on every boot of the NAS via autostart: http://www.synology-forum.de/archive/in ... 0b19d908b7
(unfortunately it is in German language but I am sure you can understand how to do it)

I am not quite sure but I think hdparm is already present on the Synology NAS by default; if not, I read it can be installed.

Perhaps this will work if for any reason the WDIDLE-utility won´t do the job corectly?

Alex

Hello Alex,

hdparm will not work with the Western Digital Green drives.
You can try on yours: hdparm -B 255 /dev/sda (255 is to deactivate). If you type hdparm -B /dev/sda, it will display the current value.
One thing to know about the hdparm is that it is not resilient, this means that you will need to pass this command at each startup of your system but a script for this is easy to setup.

Pascal

[Lex]

First results:

I edited the script like you suggested to: "/usr/syno/bin/syno_disk_ctl --wd-idle -s300"

With the command "/usr/syno/bin/syno_disk_ctl --wd-idle -g /dev/sda" I then checked the result: it showed me: "Get /dev/sdd WD idle3 timer [44] successfully"

The value of 44 looked weird; I took the disk and read out via desktop-PC: WDIDLE3 showed me "4.4 seconds". This of course is way to short. I tried with different values and found out, that "255" is the biggest value that can be set using the DiskStation.

Setting to "255" will result in "Get /dev/sdd WD idle3 timer [255] successfully".
I can not determine for sure, how long "255" actually is (in seconds), but it is long enough that the LCC is not increasing any more (my DS is running for an hour or so after making this setting now).

Conclusion: for some WD disks, disabling the timer will NOT work (as for my WD5000BEKT-22KA9)! The drive reports, that the timer is off, however this is not true and the LCC keeps increasing fast!
In this case you can try to set the timer (NOT disabling it) to a high value (the highest value seems to be 255); use this command: "/usr/syno/bin/syno_disk_ctl --wd-idle -s 255" in this script:"/usr/syno/etc/rc.d/S99WDidle3Dis.sh". (Don´t forget to shut down (restart is NOT sufficient!) the DS to let the settings take effect). This worked for me and the LCC is not increasing any more since then.

Alex

[Lex]

Second look:

my DS run over night now and I took a look at the LCC-values this morning:

for the three "00KA9"-disks the value increased by 2 (from 45 to 47) but for the "22KA9"-disk the timer increased by 223 (from 2082 to 2305). This is still much less increasment then before my modifications yesterday evening, but still to much for my taste.

Especially very interesting is the differrent behaviour of the two disk-types although the same settings have been applied! They seem to handle their timers different.

I then activated the hdparm - script and set the PowerManagement of the drives to disabled (value "255"); the drives settings where "128" before, that means "medium".

I created a "S99hdparmboot.sh" and used this script:

#!/bin/sh
/usr/syno/bin/hdparm -B255 /dev/sda
/usr/syno/bin/hdparm -B255 /dev/sdb
/usr/syno/bin/hdparm -B255 /dev/sdc
/usr/syno/bin/hdparm -B255 /dev/sdd

The script reportet, that the PowerManagement has been turned off for all four drives.

Now I will check the LCC-counter again this evening to see, if it still increases.

Of course I still want to find a timer-value for the "22KA9"-disk that leeds to the longest possible timer-setting working for this drive-type, but that probably will take hours of testing... I am not sure if it is worth it, when the hdparm-setting will do the trick.
We will see this evening

Alex