flyride
-
Posts
2,438 -
Joined
-
Last visited
-
Days Won
127
Posts posted by flyride
-
-
The goal is to keep this thread fully "mainstream" and not necessarily aligned with all the dev work that is constantly happening around RedPill. As things move from development into mainstream I intend to update, and have updated the core post.
If you want to be aligned with the very latest work, you are a dev tester and should understand the enhancements associated with development.
-
The vmdk doesn't know if it is a SCSI or SATA device, it's the virtual controller that does this.
I cannot imagine there is anything in the TCRP loader that has any bearing on it. There must be something with the install that differs from the tutorial recommendation.
Post a new thread request for help (as opposed to spamming out this tutorial thread) and include screenshots of what you are doing.
-
Is it possible you still are attached to the old vmdk instead of using the new version?
-
My understanding is that the Redpill stack internally blocks some of the Synology IP's and services involved in telemetry (thus automatic updates):
Those have nothing to do with third party SMTP services. As @Polanskiman I have no issue with email delivery from Synology, so I would point to firewall, regional filtering, or other external influence as the cause to the problem. It might be worth Wiresharking the box and filtering for SMTP to see what is going in and out.
-
Right in the link description it says this "✔ Supports FIS switches based on port multiplier switches or commands."
So yes, your card uses a port multiplier. Generally if there is a SATA device driver required for Windows, it is a port multiplier card.
-
Not a good idea for production, but you should be able to umount a shared folder and then mount something else in its place?
VEEAM >> Hyper Backup. Not sure why everyone is in love with Hyper Backup; there are much better solutions out there.
-
1
-
-
Here's the method for deleting your matched serial numbers from the DSM 7 databases:
-
1
-
1
-
-
Try copying these files out of /var/log (in case restore is needed), then delete them, and reboot
smart_quick_log
smart_extend_log
disk_overview.xml
This would solve it for sure in DSM 6, but it may be that some of this data is now database-embedded in DSM 7. Post results and I'll dig some more if necessary.
-
What DSM version is running? Is this the forum or your personal system?
-
The experimental platforms are not in the Loaders and Platforms matrix. But I would think it would be equivalent to the 1621+ which is Ryzen 4-core/8-threads, and 16 threads in the kernel. So it's probably that they are trying to do too much on the limited 2-core CPU.
-
1
-
-
7 minutes ago, WiteWulf said:
You're absolutely right: the different DSM software version that target their different hardware versions aren't limited by thread or core count. The kernel will take care of scheduling threads across however many cores you throw at your build, you just won't see it in the System Information part of the web UI, as that's hard coded to the platform type and completely ignores what hardware is actually in the device.
This is not correct. There is a finite number of threads configured by kernel parameter for each platform. These are specifically highlighted in the table.
The kernel thread limit is almost always larger than the number of physical threads on the DSM platform.
In the case of the DS3622xs+ we can use 24 threads, even though the actual hardware only offers 12 threads.
Any additional threads available with your hardware (beyond 24) will be ignored.
-
2
-
-
This tutorial is not the place to ask for or receive handholding advice on basic installs. Please post in the DSM Installation forum.
(topic that was spiraling out of control was moved)
-
1
-
-
In DSM 6, the slot positions were generically depicted to the limit of MaxDisks, regardless of how many slots were physically in the donor platform.
In DSM 7, there is a graphical representation of the platform and any additional slot positions are not shown.
Again, in the detail list of drives, all active devices are visible, present and fully operational.
-
The Synology expansion units have a completely different device connectivity structure that would be very difficult to emulate, while we generally are just adding SATA ports on top of those that are natively embedded in the platform. Bluntly, our way is simpler.
If you really wanted to make the graphic, somehow grafting the images and logic in from one of the other units would probably be the easiest way. But it would get overwritten on each upgrade, and it is, again, purely cosmetic. For most of us, not really worth the cycles.
-
1
-
-
2 hours ago, AdictiK5 said:
So there is no way I can see all my HDDs but only 4?
Drives >4 are functional. They will show up in the drive list and can be addressed normally. They just won't be displayed in the graphic.
-
If you want to, but probably not necessary. As you can see you can recover from a fresh bootloader in a matter of minutes. There is no unique configuration data stored on the bootloader, just corresponding code associated with DSM itself.
-
The bootloader is modified by DSM each time there is an update (or initial installation).
-
This card uses a port multiplier which is not supported by DSM. This is well documented.
This eBay listing even specifically calls out Synology as a limitation.
https://www.ebay.com/itm/234683011801?chn=ps&mkevt=1&mkcid=28
On 10/3/2022 at 10:03 AM, AdictiK5 said:How can I see all the hard disks graphically on the interface because at the moment I only see 4?
This is hardcoded to the DSM platform. I have not seen anyone change this yet.
-
On the surface, I would agree with your proposed command to restore the array. However I have two comments and will reiterate strongly my prior advice:
1. The sdaf disk is missing altogether from mdadm and the sdaf1 and sdaf2 partitions should be present in the md0 and md1 arrays. This is an abnormal state. If you had just dropped the disk, there should be evidence of the md0 and md1 missing array members.
2. The SHR is really, really fragmented and is a liability even if it was healthy. You will be served by rebuilding the SHR from scratch with all of the physical disks present from the beginning. I would not include the virtual disks and build a second Disk Group if you need them.
Given that your data is intact, you really MUST offload the data and get a backup. You have time to procure or borrow adequate storage. Go get it. Then delete and rebuild the SHR which resolves both problems. It would also be a safe opportunity to correct the incomplete DiskIdxMap/SataPortMap.
-
1
-
1
-
-
Because you have non-standard disk numbering (probably some problem with your DiskIdxMap/SataPortMap), you will need to re-run the discovery command as follows:
sudo fdisk -l /dev/sd* | grep "^/dev/" -
3 minutes ago, jcgechevarria said:
I configured the loadeer with only two drives to start with, then I added a third drive, and the system saw it with no problem, then I added a fourth drive and the system saw it with no problem too, and I was able to create a RAID 5 with all four drives, the thing is According to the information in the post, DS920+, "any disk ports left unpopulated at loader build time will not be accessible later." But in my case they have been seen and are accessible and usable. Another mystery??
On the other hand, as I said, I use an Intel Core i5, but in the DSM system characteristics, it indicates a Core i3... am I losing processor power or some feature?
It depends on which TCRP loader that you have used. The "stable" versions of TCRP create a static disk map which does not support dynamic recognition of other ports. Some dev versions have the ability to dynamically allocate ports as they are seen.
The GUI representation of the CPU is hard-coded to the Synology hardware of the DSM platform that you are running. It is a cosmetic error. Your CPU is still utilized up to the thread limit indicated in the matrix.
-
On 10/2/2022 at 7:35 PM, lokiki said:
I have still access to my files, only this message is display (Volume Degraded)
Before you do anything else, be smart and ensure all your files are backed up elsewhere.
Does the system not offer the ability to repair the array in the GUI? If it does, that is preferred over manual intervention.
If it does not, this command should help identify where a partition is not being serviced. Post the results.
sudo fdisk -l /dev/sd? | grep "^/dev/" -
3 hours ago, capillonpub said:
I'm a bit lost, sorry! I just put hands on a HP microserver Gen8, with an Intel Celeron G1610T (a e3-1265l is on the way for an upgrade).
I plan to install Xpenology DSM7.1 BareMetal (I had an issue with a previous install on esxi with an old N54L, the ssd was erased by high summer temperature!).
Which DSM plateform can I use? I think the GEN8 is Ivy Bridge hardware, older than Haswell so DS918+ should not be possible; then do you confirm that DS3622xs+ is the best choice in this case?
DS3622xs+ is a good option for most hardware and preferred if you don't have/don't need transcoding hardware support. I hope this is pretty clear from the ranked matrix.
-
1
-
-
AFAIK that is a normal and spurious log message.
Automated RedPill Loader (ARPL)
in Loaders
Posted
https://xpenology.com/forum/topic/61634-dsm-7x-loaders-and-platforms/
Your CPU has 56 threads. The max supported on the mainstream platforms is 24 (see the link above).
You will need to virtualize and run other workloads to maximize the CPU. If you want to run baremetal, change to one of the platforms that supports 24 threads, turn off Hyperthreading and then you can use 24 cores out of 28.