fonix232
-
Posts
324 -
Joined
-
Last visited
-
Days Won
1
Posts posted by fonix232
-
-
well it does look like you have hung hopes on 2 devs. jun and trantor for drivers.
It has always been like this. During the 4.x times, even during 5.x, we only had 2-3 devs, and ~200-300 users depending on their work.
Maybe that's because there is almost no documentation about how to make a bootloader.
If there are some good docs on how to build an image, maybe more volunteers will show up.
Uhm... That's not true.
The current loaders are nothing more than Syno's own boot solution (which you can easily reverse engineer if you buy a syno product and take a look around its inside working), paired with Grub (which has plenty of documentation), and a custom solution to get around their protection (which again, if was documented, Syno would patch in no time).
No offense, but being spoon-fed all the information is not the way to become a developer.
-
@fonix232
Interesting. There is a theoretical point about not putting more restrictive terms on GPL code but we're only be arguing about a small part of the system.
Synology is based in China. They can crap on GPL, unfortunately (theoretically even synobios should be GPL licensed, but hey, they're above the law and licensing, right? /s ).
What you say makes sense. Synology are not immune to the same market forces that have shaped the computer industry and that means first - you can only make money selling boxes for so long and then you need to unbundle the software and services.
In my understanding, there will be three different products: the hardware (which comes with software + services), software (without services), and services (only if you have software). Services would be subscription-style for software-only purchase (which is not an issue since enterprise customers won't use QuickConnect or Cloud2).
Second, when you do unbundle, if you go do the route of selling upfront software licences, then at some point you find you are not getting new users and you run out of reasons to sell new licences to your existing users.
Hence Cloud and subscription based functionality.
The services and the software would be unbundled any way, and honestly, most users are "doesn't matter what price, give me a full solution for my money". We only make a small percentage of Synology users, but the licensing is aimed at us (and also aimed at knocking companies like U-Nas out of the ring). Imagine that for 30/50$ you get all 6.x releases, all security updates included, with a bunch of drivers ready to use, and your drives a click away. Only 50$, which is pretty much half the price one would cash out for a new HDD.
However, this will only work for them if the guys who made their reputation in the company by shifting boxes can change their mindset to the new order. My experience in the industry says that is not certain.
The industry already made the statement. More and more enterprise customers are getting fed up with Syno's attitude on over-pricing hardware and dropping support. I know for a fact that Synology receives more and more requests from enterprise partners and customers for the ability to run DSM on their own hardware builds. The more this is ignored, the more customers Syno is going to lose.
-
"Please be informed that piracy is not tolerated in the forum. The use of Quickconnect IS an abuse of Synology's services." Forum Mod. who doesn't seem to get GNU.
OK I'll bite.
The fact that the Quickconnect code may be covered by the GPL does not give us the right to use any service that it connects to. We could use the code to define how to implement our own QC service and modify our version of the DSM code so that it points to ours not Synology's. But using theirs, indeed, creating false credentials in order to do so, could well fall under the scope of various computer misuse acts around the world.
This.
The code might be GNU covered (oh wait, it isn't, it's Synology's own license, just like their synobios kernel module!), but the service has an EULA, and TOS. You violate both by not using an original Synology device.
Let's be fair here. Synology makes a great product, both hardware and software. We take their software, which is a sales point already, and then we try to piggyback on their SOLD service, for free?
Just saying, that I've talked with some people from Synology. Not leadership positions, but people invested in the development of DSM and other Syno products. They confirmed that based on their current tasks, and approaches Synology is taking, it is very likely that in the near future Synology is going to license DSM separate of devices. It's mainly because of pressure from enterprise partners who wanted to run DSM on their own HW - and the first step was making the KVM version of DSM possible. It has all the hw kill switches removed, and it is possible to run that image in a regular KVM or any other VM solution, if you have the boot disk (which is eerily similar to ours, sans the model-specific stuff, plus a few VM-specific changes through the whole system). So far, based on what they got as tasks, the whole shebang will be licensed to anyone who wants to buy a license, but I'm not sure about the format ATM. But a pretty heavy-load safety feature that depended on Synology-specific HW was scrapped before the 6.1 release.
-
Synology did not kill Quickconnect. They simply disabled any old serial format DS3615xs. You can generate a new format serial, and re-register, and use it as before.
-
This week I plan to change the forum engine to IPBoard (https://invisionpower.com/)
What do you think about this idea?
It has a good bunch of issues with every version, and the support is not the best (at least that was, with older versions back in 2010-2014, don't know if it changed). Plus, it's paid. Also, PHP...
I still think that an open-source solution that is NOT PHP-based (say, any node.JS based ones) would be better for us. Maybe even with built-in chat?
-
well it does look like you have hung hopes on 2 devs. jun and trantor for drivers.
It has always been like this. During the 4.x times, even during 5.x, we only had 2-3 devs, and ~200-300 users depending on their work.
-
I could although not sure what the purpose would be. The testing was aimed at v1.02a. Let me try anyway see how that goes.
EDIT: Tried with Xpenoboot 5.2 and everything loads properly. I didn't try installing DSM as I don't see the point but it seems to work fine with Xpenoboot 5.2.
The point was to see what difference makes it not boot. So 5.2 works, meaning that kernel config is working. Did the 1.00/1.01 loaders work?
None worked. 1.00, 1.01, 1.02a and 1.02a-test
Okay, so it is definitely a difference between Jun's loader and Trantor's loader. Maybe different kernel defconfig, some feature missing?
-
I could although not sure what the purpose would be. The testing was aimed at v1.02a. Let me try anyway see how that goes.
EDIT: Tried with Xpenoboot 5.2 and everything loads properly. I didn't try installing DSM as I don't see the point but it seems to work fine with Xpenoboot 5.2.
The point was to see what difference makes it not boot. So 5.2 works, meaning that kernel config is working. Did the 1.00/1.01 loaders work?
-
I doubt it's the loader, it sounds more like a kernel panic. There's no ramoops after reboot, right?
Yes by loader I meant the kernel since it hangs while the kernel is loading. No ramoops. Only thing I can see is the usb/serial connector LED that stays on continuously instead of blinking when there is data passing in an out. It definitely hangs there.
That's a though nut to crack, especially without logs, or access to files.
Can you boot 5.2 on it, and see if it works?
-
Recalling a bug I encountered a while back with a certain motherboard, try resetting the BIOS to defaults and don't change anything except ensuring ahci is active. If that still doesn't work, remove the cd drive entirely or disable in BIOS if that option is available.
I had a similar bug with my mobo (MSI B250I PRO), but it wasn't about booting. It booted, but network was not accessible.
I had already tried a BIOS reset and that didn't change anything. I removed the ODD but still see those two lines and then nothing. I'm guessing the loader is not liking some of the hardware in that machine.
I doubt it's the loader, it sounds more like a kernel panic. There's no ramoops after reboot, right?
-
Syno Community works fine on dsm 6.x with the DS3614xs.
The packages i use are sabnzbd, couch potato, sonarr and the respected dependencies.
A good majority of the SynoCommunity packages won't work unless you install them on 5.2, and upgrade to 6.0. The reason is that SynoCommunity packages create specific users and groups, a feature that was removed in 6.0 - not entirely removed, but modified in a way that these packages need serious retouching to get them to work again. Further info on the issue here
I tried using docker n couldnt get it to work. Also i read somewhere that the docker wont be working for the near future on DX3617xs. But could try it on
Docker is quite simple to get working. Most of the packages you seem to be using are already available in the LinuxServer.io packages, ready to deploy. Think of each Docker container as an app running in its own chroot. If you need some help, I can give some pointers in private (since this thread is about the loader, not Docker or packages).
And the latter part is pure bullshit. Synology implemented Docker for a single reason - it's an industry-leading technology of containerizing stuff. I wouldn't be surprised if they replaced their apps with containers in the future.
-
I had read that the 6.x loader uses synology model that didn't support SHR raid. Is that still true? If I do a new install using xpe 6.0.2 update 7. Would I be able to use SHR .
Thanks
Synology is discontinuing SHR in favor of regular RAID arrays. Newer models and pretty much 6.x on Intel-based HW removes the option to create an SHR array, and offers regular RAID instead. However, existing SHR arrays will still work.
-
Hi, Booted a vm on ESXi 6.0 and updated to the latest (DSM 6.1-15047 Update 2).
Managed to create a disk (Virtual Only).
Was only planning on using it for a downloading client with the SynoCommunity Packages but the packaged arent compiled yet for the new DS3617xs
But looks good other than the packages issue.
Will shut it down for now untill the packages are available. Also ive not put any MAC addresses in the file and still works fine
The SynoCommunity packages do not work on 6.x, it is not an issue with the DS3617xs.
What software do you need? Most probably you'll be able to run the same via Docker (which should be the preferred method any way).
-
Recalling a bug I encountered a while back with a certain motherboard, try resetting the BIOS to defaults and don't change anything except ensuring ahci is active. If that still doesn't work, remove the cd drive entirely or disable in BIOS if that option is available.
I had a similar bug with my mobo (MSI B250I PRO), but it wasn't about booting. It booted, but network was not accessible.
-
sata0 is the HDD, sata1 is the ODD, sata2 and sata3 do not show up and most probably are empty headers as you said. sata4 is the esata port.
Yes I am aware those messages are events and not errors, yet kernel loading seems to stop and no further output is seen on the consol.
My question actually lies on why this is happening.
Doesn't the ATA port numbering in software begin at 1, while it begins at 0 on the board? Though it is possible that the controller provides 4 ports, and a separate controller is used for eSATA.
I doubt that those messages are related to the reason why it won't boot further. Can you provide full logs from start?
-
An option to what?
Cloud² seems to be a Syno offering to justify the price of their hardware, and a somewhat-competitor to Google Drive/Dropbox/OneDrive/etc. - I guess they noticed that many sync some content from their NAS to cloud solutions, and wanted a piece of that cake too.
-
it's' kernel config parameter,
CONFIG_X86_MINIMUM_CPU_FAMILY=8
Juns bootloader he creates a custom zImage which loads up the DSM software, could he, compile a new kernel with the LIMIT set to something like 24/32 something in the region of 2017... server cores. Seeing as we can get a single CPU with 8 cores then HT to 16 logical, 8 is becoming rather limiting.
Actually, the bootloader only creates a custom zImage (bzImage) for AMD boards. Otherwise the zImage is the stock 8451 zImage from 3615xs.
Compiling our own kernel is an option, though it is cleaner if we use the official one (since it gets replaced at update any way).
-
1. Stop using Quickconnect, use port forwarding and any freaking domain name. Hell, there are free ones, and DSM includes a DDNS service. Use it!
2. No, they did not ban users. They simply pulled the old DS3615xs serial format from their DB, and since it's invalid, you can't remove it.
3. The quickconnetor.orgfree generator does NOT create valid DS3615xs serials any more. If you want that so much, use this: https://xpenology.github.io/serial_gene ... r_new.html
-
I've been trying to use v1.01 and v1.02a on a laptop Toshiba Tecra I had laying around for the sake of testing. So far it has been a failure. I'm using a serial consol to see what's happening and this is what I see. There is nothing else in the output other than this:
[ 2.857140] ata3: send port disabled event [ 2.857152] ata4: send port disabled event
Looking at the bios ata4 is the esata port. There is no mention of any ata3 so not really sure what that port is. I would need to open up the case to see.
Anyhow, ideas on the above?
Most laptops come with 4 SATA ports. SATA0 is usually the CD drive, SATA1 is the main HDD, SATA2 is most probably an unsoldered header on the main board, and as you discovered, SATA3 is the eSATA.
Any way, those messages are not errors, it simply means the kernel picked up there's no device on them, and disabled them, simple as that.
-
Content won't be a problem either, migrating posts from PHPbb to something else is very easy.
A good majority of the content is not even active any more, so archiving it in a fixed format, and starting anew might help the server loads and also give us some flexibility for future threads.
-
Okay update on the troubles I'm having. Installed an Intel Dual Port 1G PCIe card into the server boots perfectly and got all configured etc... including SFP+ modules.
ran dmesg found that the Broadcom driver wasn't loading so I downloaded the Broadcom firmware
bnx2/bnx2-mips-09-6.2.1b.fw and copied the file into the /lib/firmware... directory where the other Broadcom drivers are.
Rebooted presto, new NICs show up in my Network Interfaces...
New issue when taking out the Intel Dual Port 1G PCIe card the server now doesn't boot and I can't access the DSM via port 5000, I can ping the assigned IP addresses via another machine.
I'm open to suggestions, I have 2 servers with these Broadcom NICs and only one Intel Dual Port Card (I don't need 2) I already have enough networking on these boxes each with 2x SFP+ 10G 2x RJ-45 1G, I'm really looking to be able to boot and sustain from only the Broadcom NICs
When you booted with the Intel PCIe in it, did the NAS see the Broadcom NICs?
Also, try to SSH in, maybe that'll give some information. A serial console would also help.
-
1. IP does stay the same, especially if you use the same MAC address and/or configured static IP.
2. No need for a "new" flash drive. Just use the same drive, and fill the info accordingly.
3. If you do use a new flash drive (i.e. another one), make sure you use the PID/VID of that drive, and the old one's.
4. No need to download 6.1 PAT, just use the automatic install (it will only install the main 6.1 release, not the updates)
Otherwise, yes, it is correct.
No. It is correct. I was giving MY methodology. Feel free to give yours..
1.. It didn't for me, twice - and it was a static IP address on the box previously..
2.. Do you think I don't know that? I 'chose' to use a new, spare, one (as I wanted the previous flash drive as a backup. It is in Alpha, after all)..
3.. It's an identical model of flash drive so the PID/VID are the same..
4.. I wanted to install it manually - and I will continue do so so, so I know the file that that gets installed is the one I tested..
#H
1. That's weird. Routers tend to give the same IP address out, even if it isn't in the DHCP range, as there's an active lease for the MAC-IP pair, which gets picked up. How long is your DHCP lease in your router?
2.-3. Okay, but since you did not specify this, it can be misleading for others who might be not as versed in this topic as you or me. Just wanted to clarify.
4. That's alright. Still just saying, for the completeness of your statement, that it is not required, per se, to download and install the PAT manually.
-
Is upgrading from dsm 6 (jun 1.0 loader) as follows:
- Flash the new 1.02a loader
- Take out old usb drive
- Reboot with new loader
- Choose upgrade
- Profit????
Anyone who can confirm those steps?
I've done it already, twice, on a test machine I have.. Bare metal DS3615xs..
I used a second, new, USB drive with the new loader on it..
Steps are..
- Shut down existing 6.0.2 NAS.
- Insert new USB drive flashed with new 1.02a loader. Settings should be edited as per your original loader for 6.0.x (Pid/vid/Serial number/MAC addresses/Sataportmap etc). I had to move Sataportmap command to the upper line, so it was used..
- Boot to new USB drive on your NAS. Select defaults (or just leave it).
- Scan network for new IP address (it doesnt stay the same) and open in browser (I don't use the find synology method).
- It will ask you if you want to Migrate or Clean install - choose to Migrate..
- Use (previously downloaded) pat file for 6.1 to update the box (200meg or so)..
- Box will update and reboot.. Logon again, set static IP address etc..
- Update to latest patch 2..
- Reboot.
- Done!
1. IP does stay the same, especially if you use the same MAC address and/or configured static IP.
2. No need for a "new" flash drive. Just use the same drive, and fill the info accordingly.
3. If you do use a new flash drive (i.e. another one), make sure you use the PID/VID of that drive, and the old one's.
4. No need to download 6.1 PAT, just use the automatic install (it will only install the main 6.1 release, not the updates)
Otherwise, yes, it is correct.
-
I am holding off. I do not have the need to go to 6.1 yet. Also, I am not liking those "volume crashed" or whatever messages that appear on screen even if they seem to have disappeared with update 2. Finally there are still some module loading issues with his release and also still no AMD support.
Only Jun himself can answer that question but I suppose that if he has the time he is probably also looking into the recently published source code of DSM 6.1
Since Update 2, my NAS does not produce the 12h volume error, or the inaccessible DSM dashboard. I'm currently at 1d 6h uptime, mainly because I installed a 5.25" -> 4x 2.5" bracket last evening. But apart from that, my NAS has been running non-stop since last Friday, with Update 2.
3615xs generated serial registration problem
in Packages & DSM Features
Posted
For one, as above, stop using QuickConnect.
For two, yes, it is a well-known "issue" that Synology changed to a new serial format ~2 years ago. Not much that you can do about it, except maybe Google a bit, because there's already a generator for the new format.