Why releasing firmware updates should be done often

I have developed, maintained and sold as product manager many devices. Many devices with firmware, like Routers, Mediaplayers but also NAS Devices.
you can say that I am expert on various levels (from Technical point of view, but also from Marketing / Commercial side making the right choices for the company and customers).
One of the latest products I created from idea to end product (of course with the help of other people) was the a very nice NAS available on the market today.
it won around 7 prices (WOW!, we actually did something good).
Unfortunately I am not involved in that project anymore as I am no longer employed, but I am still following the process of this product, as I think this product is still awesome, with Private Cloud (with Tonido) and the Marvell 6282 chipset .. (Best features, performance and quality for the price you are paying).

But every product has bugs and my idea is always to test the device on main topics to see if a firmware upgrade can be done safely and if data will not affected. If so, check the release notes and test the fixed bugs or new items inside. Release the firmware as a beta firmware and have it tested further by your userbase. Make a single point of contact and collect the input of the users. of course it will take time, but the buzz about your support and device will be granted with people advising your product to other customers/users.

When people are postive about the beta, make this beta public and go further with another beta.

I have seen great success with the Conceptronic CH3SNAS and CH3MNAS where I released many fw, but also for the Conceptronic CFULLHDMA(i) where I released almost every week a new firmware. (I had around 700 beta testers, and from those 700 around 50 gave me daily/weekly input of their tests.. Awesome. (not only for me, but also for the factory). I can say that due to this, more devices have been sold than if we would not have done it.
unfortunately above 3 devices are not sold anymore nor maintained by Conceptronic or manufacturing factory. As long as you buy in good quantities a factory is willing to maintain it, or you pay a nice NRE fee and they run for you as well ;-)

Your keyusers will love this way of working (you show development / progress). And Less experienced users will not even try to upgrade, but will do this only when they encounter an issue or have called with your local service desk and are requested to upgrade safely to a released version.

I encountered an issue at my side some time ago: I was unable to store more than 88% of the total disksize of the NAS. After 88% copying of files to the NAS was giving a lot of troubles.. which resulted in broken and failing copy actions. Bummer …
I do not know if there are more users having this issue cause it might only happen when you store a lot of data and files (I was using a 6TB RAID0 configuration)
so to store more than 4.5TB will take some time ;)

I deleted a lot of files and I was able to store data again.. I forgot about this issue for some months but I saw it again when I tried to copy 2.5TB of data last week. Checking the website: no new fw release. ??? That’s wierd cause I expected a new one as I know that there are some things what could be improved.

So this time I was not that happy and took some time to analyze the issue. It looks like it was related to samba, but after debugging the logfiles there was no real error what could be related that samba was failing.

To make a story short: I found out that the kernel version the NAS was using did not optimze the filesystem nodes properly. So a lot of space was wasted when there came data in or out. As a result of this there was no space anymore to store more nodes when we store more data in the system.
I managed to update the kernel (do not ask me how)…(Actually you need an ssh enabled fw release (not available to the public), the official fw release did not enable ssh (removed). No idea if a released GNU/GPL sourcecode will give you the opportunity to upgrade the kernel image.

for me by only upgrading the kernel:

1. My 88% issue was fixed, I can store data again and as a positive side affect:
2. Improvement of write speed towards the NAS: the device reached a 40MB/s write speed what I never saw before. Finally the Marvell chip could be used what it was designed for… Storing at this speed was not seen before. (Read speed around 50MB/s but I have seen a 65MB/s, which was the top, but could not be maintained).
very very satisfactory.

I wish they will release a fw update soon .. because unfortunately it seems that no new firmware upgrade has been released yet. I do not know why, and actually I do not want to know why …
but I think that their policy is not to release firmware upgrades often and first test it internally.
A totally different strategy than what I should have proposed or done. As single person a device like a NAS has too many features you can test in a timely matter and in various configuration settings. You need the help of others.

People might disagree about releasing firmware upgrades often.. If so: I would love to hear the comments why a vendor should not release (beta) firmware often.

Note: I have not mentioned names of the NAS device nor I have mentioned brands and factory names as I think this is not important for this story.

Trying to understand: CH3SNAS & NFS = Help needed !

I’ve written earlier a simple howto for users with a CH3SNAS to work with NFS. UNFS is a user space implementation of the NFSv3 protocol. This is working flawlessly on the CH3SNAS with various Linux PC’s I am running (CentOS, Debian, Linux Embedded device etc.. )

but for a VMware implementation I need the READDIRPLUS procedure which is not available in UNFS. So I need to run Kernel-NFS on the CH3SNAS

Fonz fun_plug is providing Kernel NFS (enabled in the CH3SNAS) and UNFS. Installation is easy and almost fool proof.

The issue is following: on all my linux boxes (CentOS, Debian etc.. ) I get following message:

mount: mount to NFS server ‘192.168.1.60’ failed: timed out (retrying).

  • rcpinfo -p 192.168.1.60 gives all correct info
  • showmount -e 192.168.1.60 gives all correct exports
  • rpcinfo -u 192.168.1.60 gives:

program 100005 version 1 ready and waiting
program 100005 version 2 ready and waiting
program 100005 version 3 ready and waiting

  • Showmount -a 192.168.1.60 gives the mount points I ‘tried’ to use:

All mount points on 192.168.1.60:
192.168.1.100:/mnt/HD_a2

and even syslogd gives on the NFS server (CH3SNAS):

Jun 12 18:33:07 DEVELOPMENT daemon.notice mountd[1868]: authenticated mount request from 192.168.1.100:615 for /mnt/HD_a2 (/mnt/HD_a2)

  • Portmap is running at server and client!!

I am totally not understanding why I get time-outs. I do not have a firewall, I do not use iptables. Today I tried in another environment the same: always the error:

mount: mount to NFS server ‘192.168.1.60’ failed: timed out (retrying).

suddenly here at home after about 20 minutes trying to mount the NFS share, the NFS server has been mounted on my local system. But when umounting and mounting again I have again same situation. In my work environment it was totally not possible (standard Debian setup and a CentOS configuration) to have the NFS share mounted. Although it is similar setup ..

So although al tools tells me that everything is correct: I am not getting any mount :-(

Is there anyone with enough NFS knowledge who can tell me why the NFS share is not mounted?

RDC 2891 chipset LANDISK

For those people with a RDC 2891 chipset LANDISK (low-cost)

I have managed to get a new kernel driver: kernel-s2891-010b1 

There are no release-notes available yet, but I know for sure that a number of stability / compability issues are or will be solved. Just upgrade the file through the webinterface. Note: No warranties.

Note: this is NOT a kernel update for the RDC2882 chipset. Only for the RDC2891 chipset

Change-Notes

Fixed: Mac OS 10.5 support

Fixed: Mac OS 10.4.x support multi-folders.

Fixed: FileSystem Longname refine.

Add: UPNPAV support iRadio.

Add: HD volume set from web.

Message to Self: CentOS 4.4 & Bluequartz

Onboard SIS900 network card issue with Centos 4.4 & Bluequartz and no network activity available even when all config info says it is (with Linux version 2.6.9-42.0.2.EL). Problem found in disabled APM / ACPI. When disabled the 2.6.9-42.0.2.EL will not be able to communicate through the network card. When enabling APM / ACPI it is.

This error was not seen when using Linux version 2.6.9-22.0.2.EL 

I removed all kernels in between and now CentOS together with my BlueQuartz services is running fine. Even when I will have a power shutdown again ;-) Enig klein probleem wat ik heel even tegen kwam, waren de 2 volgende meldingen:

Sep  6 19:26:49 wb kernel: CPU0: Running in modulated clock mode
Sep  6 19:26:54 wb kernel: CPU0: Temperature above threshold

Changed some BIOS settings this morning. Info of temperature is not measured. Maybe that’s a solution