Develop and refine the DS3622xs+ loader

[pocopico]

20 minutes ago, altas said:

is this already in the loader visible 7.01 u2 ??

 

No, someone should add the config and the kernel patch in their repo. e.g. Dogodefi. 

Edited February 18, 2022 by pocopico

[p33ps]

51 minutes ago, altas said:

shutdown is now working on DS3622XS+ on HP GEN8 Micro

 

Hey, can you please explain what build did you use to get it working in HP Gen8?

[Orphée]

10 minutes ago, p33ps said:

 

Hey, can you please explain what build did you use to get it working in HP Gen8?

if baremetal just add this extension to the loader and also update the loader with ./rploader.sh update now :

https://raw.githubusercontent.com/jimmyGALLAND/redpill-ext/master/acpid/rpext-index.json

 

if VM, just update and rebuild : ./rploader.sh update now

 

Edited February 18, 2022 by Orphée

[Orphée]

39 minutes ago, pocopico said:

 

No, someone should add the config and the kernel patch in their repo. e.g. Dogodefi. 

To be honest I don't even know if Dogodefi is here on Xpenology board... is it @yanjun ?

Edited February 18, 2022 by Orphée

[yanjun]

2 minutes ago, Orphée said:

To be honest I don't even know if Dogodefi is here on Xpenology board... is it @yanjun ?

what's mean Xpenology board?

[Orphée]

Here, the forum, are you Dogodefi ?

Edited February 18, 2022 by Orphée

[yanjun]

1 minute ago, Orphée said:

Here, the forum, are you Dogodefi ?

guess😀

[Orphée]

1 minute ago, yanjun said:

guess😀

Thanks for your work then

[altas]

31 minutes ago, p33ps said:

 

Hey, can you please explain what build did you use to get it working in HP Gen8?

use the Tinyloader to create the Image.
its not needed to add the extension for acpid, mine was working without.

[Orphée]

18 minutes ago, altas said:

use the Tinyloader to create the Image.
its not needed to add the extension for acpid, mine was working without.

Of course it works, but if you want the physical button able to shutdown on simple click (as baremetal) you need the ext.

[altas]

ah ok  need to test
i only had the shutdown/Reboot stuck in DSM for Barmetal and Gen8

 

will give it a try later

[yanjun]

Hi everyone, I am currently trying to solve the following problem (important and urgent):

 

1, On my machine, once the data scrub is run manually, the btrfs scrub start of the first step takes roughly a few hours and works fine, while the second step of resyncing the array causes a memory overflow after roughly 10 minutes (I give dsm allocated 38G of memory). After that, it will cause a restart. I tested it twice. Although there is no data overflow, it has laid a hidden danger for data recovery after an error occurs. For detailed logs, see messages-oom-killer.txt

 

2. When you mix sata controller and mpt3sas-driven LSI card, you can't mix sata controller's hard disk and lsi card's hard disk to create raid1, normally not many people will create raid1, so I need to clarify Now, this bug will cause the startup of md0 to fail. Once you mix the two hard disks, it will fail to enter junior mode when the system starts for the second time. The phenomenon on the webui is to prompt you to install the system. I personally think that there is still a problem with the mpt3sas driver.

messages-oom-killer.txt

 

 

@IG-88 @buggy25200 @pocopico 

[flyride]

58 minutes ago, yanjun said:

normally not many people will create raid1

 

The default behavior of DSM is to RAID1 over all available devices for /dev/md0 (Linux OS) and /dev/md1 (Linux swap).  I assume we all know this but just to clarify.

[yanjun]

3 minutes ago, flyride said:

 

The default behavior of DSM is to RAID1 over all available devices for /dev/md0 (Linux OS) and /dev/md1 (Linux swap).  I assume we all know this but just to clarify.

Yes, I also observed that each HDD is in RAID1 of md0 and md1, but if the mpt3sas-driven HDD and the ahci-driven HDD are rebooted, the system will not boot normally and enter install mode.

 

 

[yanjun]

1 hour ago, yanjun said:

2. When you mix sata controller and mpt3sas-driven LSI card, you can't mix sata controller's hard disk and lsi card's hard disk to create raid1, normally not many people will create raid1, so I need to clarify Now, this bug will cause the startup of md0 to fail. Once you mix the two hard disks, it will fail to enter junior mode when the system starts for the second time. The phenomenon on the webui is to prompt you to install the system. I personally think that there is still a problem with the mpt3sas driver.

Good New:  this issue has been fixed.

I tried to replace the source code of mpt3sas, which is from the driver download page of Broadcom's latest device. https://www.broadcom.com/products/storage/host-bus-adapters/sas-nvme-9500-16i Validated on both of my LSI devices (SAS2308 PCI-Express Fusion-MPT SAS-2 and SAS3416 Fusion-MPT Tri-Mode I/O Controller Chip (IOC)). The packaged mpt3sas driver module address is as follows.

https://github.com/dogodefi/mpt3sas/blob/offical/rpext-index.json

 

@pocopico hope you can refer to my experience.

@IG-88 Hope to hear your advice on this issue.

[pocopico]

15 minutes ago, yanjun said:

Good New:  this issue has been fixed.

I tried to replace the source code of mpt3sas, which is from the driver download page of Broadcom's latest device. https://www.broadcom.com/products/storage/host-bus-adapters/sas-nvme-9500-16i Validated on both of my LSI devices (SAS2308 PCI-Express Fusion-MPT SAS-2 and SAS3416 Fusion-MPT Tri-Mode I/O Controller Chip (IOC)). The packaged mpt3sas driver module address is as follows.

https://github.com/dogodefi/mpt3sas/blob/offical/rpext-index.json

 

@pocopico hope you can refer to my experience.

@IG-88 Hope to hear your advice on this issue.

 

 

So what was the solution ? Was it related to the mpt3sas version of the driver ? Was this related to your HBA firmware level ? 

 

[yanjun]

2 minutes ago, pocopico said:

 

 

So what was the solution ? Was it related to the mpt3sas version of the driver ? Was this related to your HBA firmware level ? 

 

I personally think that the Drive Version of MPT3SAS is related to my two models of LSI devices, there is the same ISSUE. My solution is to update the driver, use MPT3SAS-40.00.00.00-src.tar.gz, included in Linux_Driver-rhel7-8_sles12-15_gen35_phase_21.0_nvme.

[cferra]

2 hours ago, yanjun said:

Hi everyone, I am currently trying to solve the following problem (important and urgent):

 

1, On my machine, once the data scrub is run manually, the btrfs scrub start of the first step takes roughly a few hours and works fine, while the second step of resyncing the array causes a memory overflow after roughly 10 minutes (I give dsm allocated 38G of memory). After that, it will cause a restart. I tested it twice. Although there is no data overflow, it has laid a hidden danger for data recovery after an error occurs. For detailed logs, see messages-oom-killer.txt

 

2. When you mix sata controller and mpt3sas-driven LSI card, you can't mix sata controller's hard disk and lsi card's hard disk to create raid1, normally not many people will create raid1, so I need to clarify Now, this bug will cause the startup of md0 to fail. Once you mix the two hard disks, it will fail to enter junior mode when the system starts for the second time. The phenomenon on the webui is to prompt you to install the system. I personally think that there is still a problem with the mpt3sas driver.

messages-oom-killer.txt 160.86 kB · 1 download

 

 

@IG-88 @buggy25200 @pocopico 

I think the same happens with mpt3sas on 918+ as well. I installed and then immediately got a critical os error on first boot and it prompts me to put in a support ticket - no option to repair system partition - after that when i reboot it wants to install DSM again. 

Edited February 18, 2022 by cferra

[yanjun]

1 minute ago, cferra said:

I think the same happens with mpt3sas on 918+ as well. I installed and then immediately got a critical os error on first boot and it prompts me to put in a support ticket - after that when i reboot it wants to install DSM again. 

If you need, I can make 918 mpt3sas MPT3SAS-40.00.00.00 extension for you.

[cferra]

I think it would be good to do as a driver pack - i think I want to migrate to the 22xs - i couldn’t get the drivers to work at all with tc and 3617 so i tried with 918, they worked there but I ran into that issue.  
 

for right now I had to fall back to jun’s 6.3.2 3617 

[cferra]

While I have no data on the array I have some ability to experiment too so if I could be useful in that regard let me know. 
 

i have a x10srl-f board with a v4 xeon and an lsi 9305-24i card. 

[yanjun]

13 minutes ago, cferra said:

While I have no data on the array I have some ability to experiment too so if I could be useful in that regard let me know. 
 

i have a x10srl-f board with a v4 xeon and an lsi 9305-24i card. 

Whether you want to test 3622xs + or 918+, just convert the address of the MPT3SAS driver package as follows:

 

https://raw.githubusercontent.com/dogodefi/mpt3sas/offical/rpext-index.json

[cferra]

Sounds good!  Do you know off hand if there are mellanox driver for 22xs?

[pocopico]

26 minutes ago, cferra said:

While I have no data on the array I have some ability to experiment too so if I could be useful in that regard let me know. 
 

i have a x10srl-f board with a v4 xeon and an lsi 9305-24i card. 

 

@cferra, @yanjun I've updated 918,3617,3622 with the YanJun provided sources 

[cferra]

4 minutes ago, cferra said:

Sounds good!  Do you know off hand if there are mellanox driver for 22xs?

Looks like pocopicos repo has mlx drivers for 22xs. Awesome