Optimizing OS3 and Getting some Sleep

Legacy Products My Cloud OS 3
1

I wouldn’t really recommend doing this since Bill_s will inadvertantly say “It will void your warranty”

edit: The two clouds are blinking their slow blue lights at me at this moment. I don’t know how fitfully they slept, but I have a feeling that the cron has at least woke them up yesterday night; but they are sleeping at the moment.

edit: emerson.clarke has posted up “How to permanently disable wdmcserverd and wdphotodbmergerd”. This is an excellent post as it even has a patch to fix the capacity display.

Nevertheless, no matter how you disable the scans, this is now one of the standard user procedures that goes hand in hand with updating FW followed by disable scans.

Thus… SSH into your device and just kill the three processes

/etc/init.d/wdmcserverd stop

/etc/init.d/wdphotodbmergerd stop

/etc/init.d/restsdk-serverd stop

Yes, if you stop the first two, you probably won’t see your nice photo and video collection in your mobile phone app, but the cloud app will still work, you just have to navigate the folders yourself without search or thumbnails.

The last one is a must and it is from rac8006 at this post see user Rac8006’s post on page 13to resolve the log filling up with restsdk messages.

There are many more processes we can kill, carried over from previous versions; they are all here, nothing has changed really but for now these three should make your cloud feel brand new again.

Over the next few days, I’ll see if I can re-gain my old sleep time of multiple days. Probably Rac8006 will beat me to it, if so, please post up Rac8006.

edit: I don’t know if this helps or not, but it doesn’t hurt to use it. It was another one of Rac8006 discovery in the last firmware version.

mount -o remount,noatime,nodiratime /dev/root /

As Rac8006 says below in another post, I’m getting the 5 and 8 second wakeups, at least 3 in succession, so I don’t know what is up with that. Also rather than simplifying the Cron, it seems that OS 3 has added more jobs to the system.

Anyways… I’ll always keep editing this first post as I discover more info. The next process I might stop is the cron.

If you want to stop it now…

/etc/init.d/cron stop

This will get rid of the 3AM wakeup…

1 Like
2

I still don’t know what is the restsdk-serverd process for ? I don’t remenber it before OS3 (or maybe it didn’t cause any issues)

3

tzalewski wrote:

I still don’t know what is the restsdk-serverd process for ? I don’t remenber it before OS3 (or maybe it didn’t cause any issues)

Not sure. But I stopped it and still the device does not sleep. I checked the user.log and saw nothing in their with restsdk using grep, when that process was running. So far that’s my only issue - device does not sleep consistently and no activity.

4

Raphael

So far I can’t get the device to sleep like it did in the past.  I found that in version 303,307,308.  They added three new

cron jobs.  ga_cron_d not sure what it does.  Seems to look for a file in /Cachevolume.  It runs every day at 3AM.  But none are there.  Also ssl_cert_update.  This probably does what one would think it does by the name.  It runs a script at 0:30AM.  Then they added wdlog.  I’m not that familiar with cron.  These three files look like a crontab file.  wdlog runs 5 scripts.  4 of them at 3:01AM and one every 8 hours one minute after the hour.  I see these wake up times in my sleeptime.sh output.

I’m also seeing the 8 second wake up.  I’ve made a couple of changes and its not 5 or 6 second wake up.  I’ve tried changing the sleep time after the sync’s.  But no matter what time I sleep 10,20,30 or 60 seconds it still wakes up in 5-6 seconds or 7-8 seconds with out the new changes.  I think the change from 8 seconds to 5 seconds is because I found that the monitorio.sh script does a run-parts before getting the time.  This run-parts runs the scripts in /etc/nas/wakeup.d.

RAC

1 Like
5

rac8006 wrote:

Raphael

 

So far I can’t get the device to sleep like it did in the past.  I found that in version 303,307,308.  They added three new

cron jobs.  ga_cron_d not sure what it does.  Seems to look for a file in /Cachevolume.  It runs every day at 3AM.  But none are there.  Also ssl_cert_update.  This probably does what one would think it does by the name.  It runs a script at 0:30AM.  Then they added wdlog.  I’m not that familiar with cron.  These three files look like a crontab file.  wdlog runs 5 scripts.  4 of them at 3:01AM and one every 8 hours one minute after the hour.  I see these wake up times in my sleeptime.sh output.

 

I’m also seeing the 8 second wake up.  I’ve made a couple of changes and its not 5 or 6 second wake up.  I’ve tried changing the sleep time after the sync’s.  But no matter what time I sleep 10,20,30 or 60 seconds it still wakes up in 5-6 seconds or 7-8 seconds with out the new changes.  I think the change from 8 seconds to 5 seconds is because I found that the monitorio.sh script does a run-parts before getting the time.  This run-parts runs the scripts in /etc/nas/wakeup.d.

 

 

RAC

now you tell me, after I upgraded :frowning:  you should have said Ralphael don’t upgrade or you will lose sleep!!

6

Rahpael

For the last 34 hours my system has slept 21.  Not great but better than the past.

RAC

1 Like
7

rac8006 wrote:

Rahpael

 

For the last 34 hours my system has slept 21.  Not great but better than the past.

 

RAC

that is good enough… and mine is sleeping too at the moment with minimum process killed (only the three from above) and the remount. Will probably kill the cron too and that should do it for now…

8

@Ralphael @rac8006 Any new findings / tips ?

9

So far the best thing I found is to mount the root drive using the following command:
mount -o remount,noatime,nodiratime /dev/root /

RAC