Slow performance Firmware 4.00.00-607

after update, super slow performance transfering files

tried this on w7, w8.1 and ubuntu 14.04.

when i start transfering files it gets to 100mb transfer rates, but then it slows down rapidly. causing the transfer to take a lot more time that what it should take

in previous firmware i can transfer files to a rate of 30mb

Dashboard also superslow to refresh

How long since the update? Mine took a couple of days of indexing whatever until getting back to normal.

My WDcloud did not start to work good even after 2 weeks. It is a disaster from WD. They schould give us new firmware without bugs. But they do not do anything. I have contacted them but no help. It is the last time i’ve purchased anything from WD.

updated last saturday

GeminiMichael wrote:

How long since the update? Mine took a couple of days of indexing whatever until getting back to normal.

i just opened a ticket about this, apparently we are not the only ones with issues with this terrible update

emero wrote:

My WDcloud did not start to work good even after 2 weeks. It is a disaster from WD. They schould give us new firmware without bugs. But they do not do anything. I have contacted them but no help. It is the last time i’ve purchased anything from WD.

I read in the forums reporting of slow speed after update to firmware 4.0, but I don’t think the problem is the new firmware version. I’m currently stuck at firmware 3.0 because of other reasons and the drive suffers of a severe performance problem as soon as:

  • you start copying many files on it
  • you attach an USB drive with many files in it

In fact, the problem is that the drive is running some background processes to scan the new content on the background, even if I disable all the “multimedia” scanning and streaming features in the drive settings (and for all shared folders). These proceses consume a lot of CPU and RAM and the drive becomes so busy that it will either transfer files at low speeds or even become unaccessible until the scanning process has finished.

This problem is most serious when you attach a USB drive with a lot of files in it: in my tests, the My Cloud became so unresponsive that it was not accessible in any way (web UI, SSH, etc.) during the scanning process (for instance, it’s in this situation in this exact right moment and it has been doing this for hours now… I will leave it all night long and see tomorrow…) and before moving video files from the USB drive to the My Cloud itself using a third PC, the scanning process on USB drive connection even lead to out-of-memory errors that caused the Linux kernel to kill some processes, like the apache2 process… the result was a dead scanning process and a My Cloud drive accessible only via SSH.

There must be some memory leak, because before the crash I see the swap space constantly grow until the OOM happens.

I think WD should really do something for this. I read a FAQ on their support site that says that this problem should have been already fixed, but it is not. And I’m seeing this in the latest 3.x firmware too, so I don’t think it’s 4.x specific. Maybe you’re seeing this only now because the firmware updated caused the scanning process to rebuild it’s database from scratch.

For me, I would prefer not to have any background task, unless I request it. Or, at least, this task should be at very low priority and with a minimal footprint, at the cost of increasing the scanning processing time (but without affecting the drive performance!).

mauromol wrote:> This problem is most serious when you attach a USB drive with a lot of files in it: in my tests, the My Cloud became so unresponsive that it was not accessible in any way (web UI, SSH, etc.) during the scanning process (for instance, it’s in this situation in this exact right moment and it has been doing this for hours now… I will leave it all night long and see tomorrow…) and before moving video files from the USB drive to the My Cloud itself using a third PC, the scanning process on USB drive connection even lead to out-of-memory errors that caused the Linux kernel to kill some processes, like the apache2 process… the result was a dead scanning process and a My Cloud drive accessible only via SSH.

There must be some memory leak, because before the crash I see the swap space constantly grow until the OOM happens.

I left the system there for about two whole days after connecting a 3TB USB NTFS hard disk to the MyCloud. During this period, the MyCloud unit was inaccessible: trying to connect via SSH caused the SSH client to hang undefinely, the web UI did not respond (connection timeout), no other way to know what the drive was doing. However, a continuous disk activity was in progress (I could hear the noise of the disk constantly seeking).

At last, I had to unplug the power cable, disconnect the USB drive and turn on the MyCloud again without the USB drive being connected. As soon as I plugged in the USB drive, the same thing happened again. This even if I have multimedia server disabled and, being very quick to operate before the MyCloud got unresponsive, I could disable multimedia content for the shared folder corresponding to the USB drive.

Fact is that there are some background processes in the MyCloud that still scan continuously the contents of the drive and of any connected USB drive even if all the multimedia features are disabled. In my case the USB drive had some folders used for old backups with tens of thousands files and this seemed to make the MyCloud crazy.

At last, I found the solution:

  • make sure the USB drive is not connected
  • start MyCloud
  • connect to MyCloud via SSH
  • stop the wdnotifierd service: service wdnotifierd stop
  • make the change permanent (that is: let you start that service manually only when you want it, instead of automatically at system statup): update-rc.d wdnotifierd disable
  • plug the USB drive

In this way you stop the process which is causing memory leaks and undefinite disk I/O when you plug a USB drive (and not only). I really don’t know what’s wdnotifierd service for… I didn’t notice any lack of functionality in my MyCloud.

Should you ever need to revert that change (i.e.: make wdnotifierd start automatically on system start), login to SSH and type: update-rc.d wdnotifierd enable

 

There are at least another couple of services that scan the drive in the background and may slow down transfers: wdphotodbmergerd (which I guess builds picture previews) and wdmcserverd (which I don’t know what’s for either). You may try to stop them with the service command like above to see how transfer speeds go and then decide to disable the automatic start if you see you can live without.

 

This applies to firmware 3.x, but it’s likely it will work for 4.x too.

Mauro