IG-88

did you test to not match the usb vid/pid as i wrote above?

as you are using 918+ please keep in mind that its default config will only support 2 nic's you should change synoinfo.conf before adding a card with two more ports mentioned here: https://xpenology.com/forum/topic/21663-driver-extension-jun-103b104b-for-dsm622-for-3615xs-3617xs-918/ -> https://xpenology.com/forum/topic/12679-progress-of-62-loader/?do=findComment&comment=92682

yes, as the start process looks for update files and we dont want the update to be installed i would remove the files they are clearly belonging to the update process and doe not relate to a normal running dsm system

live linux booten (der dsm loader verändert was man be "lsusb" sieht) oder den stick an einem anderen pc z.b. mit windows prüfen (ist im tutorial für 6.1 auch beschrieben, neben dem geräte manager gehen auch USBDeview und UsbTreeView) wenn der kernel ud treiber unterschiedlich sind dann stimmt was mit dem loader nicht ramdisk.lzma und zImage (kernel) müssen eigentlich zur extra.lzma passen die die treiber enthält generell wir dann etwas nicht geladen (vermutlich treiber) theoretisch schon aber boote doch einfach ein live linux von usb oder nimm eine platte auf der bereits open media vault installiert ist und starte davon (letzteres ist ja auch ein nas system und du hast eine webgui um freigaben anzulagen so das du mit windows an deine daten kommst (was die meisten wollen) aber ein update auf 6.1 verändert nichts an den daten auf der raid partition, das fasst eigentlich nur die system partition, wenn du einen usb mit 6.1 fertig machst und der im netzwerk gefunden wird sollte es genügen einfach eine migration auf 6.1 durchzufügren indem man das passende *.pat file installiert

ok, i know i don#t write in my native language and i repeat myself (3rd time writing this) - when it boots and you find it in network the cpu can be excluded, if the cpu would be to old then it would not boot at all because the kernel would not start, no kernel, no drivers, not found in network the positive proof that its found in network is enough to for the conclusion that that the cpu is fine for 1.04b 918+ the next point for you (see above) was to deliberately change the usb vid/pid to not match the real values and test if you can install without error 13 quick glance at the dmesg, there are 4 + 6 sata ports found and a lsi sas controller would only be left with 2 ports usable as 3615/17 default config is max. 12 disks - but there does not seem to be such controller, mpt drivers are loaded (raid and non raid) without showing that a valid hardware is found [Tue Apr 27 13:33:54 2021] pci 0000:01:00.0: [1000:005f] type 00 class 0x010400 thats a lsi MegaRAID SAS-3 3008 [Fury] https://pci-ids.ucw.cz/read/PC/1000/005f ... [Tue Apr 27 19:49:24 2021] ahci 0000:00:11.4: AHCI 0001.0300 32 slots 4 ports 6 Gbps 0xf impl SATA mode ... [Tue Apr 27 19:49:24 2021] ahci 0000:00:1f.2: AHCI 0001.0300 32 slots 6 ports 6 Gbps 0x3f impl SATA mode ... [Tue Apr 27 19:49:28 2021] megaraid cmm: 2.20.2.7 (Release Date: Sun Jul 16 00:01:03 EST 2006) [Tue Apr 27 19:49:28 2021] megaraid: 2.20.5.1 (Release Date: Thu Nov 16 15:32:35 EST 2006) [Tue Apr 27 19:49:28 2021] Fusion MPT base driver 3.04.20 [Tue Apr 27 19:49:28 2021] Copyright (c) 1999-2008 LSI Corporation [Tue Apr 27 19:49:28 2021] Fusion MPT SPI Host driver 3.04.20 [Tue Apr 27 19:49:28 2021] Fusion MPT SAS Host driver 3.04.20 [Tue Apr 27 19:49:28 2021] Fusion MPT misc device (ioctl) driver 3.04.20 [Tue Apr 27 19:49:28 2021] mptctl: Registered with Fusion MPT base driver [Tue Apr 27 19:49:28 2021] mptctl: /dev/mptctl @ (major,minor=10,220) [Tue Apr 27 19:49:28 2021] megasas: 06.506.00.00-rc1 Sat. Feb. 9 17:00:00 PDT 2013 [Tue Apr 27 19:49:28 2021] mpt2sas version 20.00.00.00 loaded [Tue Apr 27 19:49:28 2021] mpt3sas version 22.00.02.00 loaded ... i guess jun used the kernel default drivers and 1000:005f is not supported in kernel 3.10 but is in 4.4 (918+) so 918+ detects the controller and 3617 does not its "megaraid_sas.ko" - thats a raid driver so you controller is in wrong mode aka IR mode not IT firmware its IR firmware and you would need to reflash it to IT firmware AND we already had this covered https://xpenology.com/forum/topic/28321-driver-extension-jun-103b104b-for-dsm623-for-918-3615xs-3617xs/?do=findComment&comment=202009 if you reflash to IT firmware the better/newer mpt3sas.ko would kick in and detect the controller

https://xpenology.com/forum/topic/29872-tutorial-mount-boot-stick-partitions-in-windows-edit-grubcfg-add-extralzma/ gibts sogar in deutsch, sollte im deutschen bereich ganz oben zu sehen sein https://xpenology.com/forum/topic/29670-howto-partitionen-des-bootsticks-in-windows-mountenbearbeiten-grubcfg-extralzma/ VERSION sieht aus normales 6.02, das wärer dann das *.pat file dazu https://global.download.synology.com/download/DSM/release/6.0.2/8451/DSM_DS3615xs_8451.pat 0000:00:1f.2 Class 0106: Device 8086:8c82 Subsystem: Device 1849:8c82 Kernel driver in use: ahci ... 0000:02:00.0 Class 0200: Device 10ec:8168 (rev 11) Subsystem: Device 1849:8168 Kernel driver in use: r8168 der storage treiber ist ahci und netzwerk ist der übliche realtek, sollte also auch mit 6.1 und 6.2 laufen das raid sieht doch erst mal ok aus, drei platten, das raid 5 sagt drei von drei platten gefunden du kannst ja auch mal loader 1.02b für 3615 versuchen das system auf 6.1 zu bringen https://global.download.synology.com/download/DSM/release/6.1.7/15284/DSM_DS3615xs_15284.pat laut der dmesg ist das eine 4th gen haswell cpu (Celeron G1840), da könntest du sogar loader 1.04b für 918+ nehmen (6.2.3 wöre hir max. version, die neue 6.2.4 läuft nicht) sowohl 1.02b 3615 als auch 1.04b 918+ laufen mit uefi und csm mode, beide loader sollten ohne größere probleme anlaufen und im netz zu finden sein https://xpenology.com/forum/topic/13333-tutorialreference-6x-loaders-and-platforms/ wichtig bei jun's laoder ist halt die usb vid/pid in der grub.cfg korrekt einzutragen https://xpenology.com/forum/topic/7507-tutorial-installmigrate-dsm-52-to-602-juns-loader/ https://xpenology.com/forum/topic/7973-tutorial-installmigrate-dsm-52-to-61x-juns-loader/ (für 6.2 ist alles identisch wie bei 6.1 nur das laoder 1.03b bzw. 1.04b und dsm 6.2 benutzt werden)

you cpu is already 4th gen (haswell) and error 13 is not related to cpu, the cpu thing is about: is it booting (starting the kernel) or not, if you can find it in network the kernel came up and loaded drivers sadly your dmesg file is not what we are looking for, it looks like a screenshot from a ssh session, there are only about 10 lines of the dmesg you need to copy the dmesg file itself to a location like u usb that dsm will mount automatically and after that you can eject the usb by thwe webgui and use it on you computer to attach it here usb drive is usually /volumeUSB1/usbshare/ , that would be the destination when copying the file it is supported the chip is pretty old and is usually working with any drivers, there are problems here and there when it comes to performance but usually its at least working with a 10th gen cpu you should use 1.04b 918+ 3615/17 is more tricky as it needs csm mode in uefi bios and you need to boot from the non-uefi usb boot device 918+ can do uefi and csm https://xpenology.com/forum/topic/13333-tutorialreference-6x-loaders-and-platforms/ if 3615/17 does not boot the right way (csm mode and the right usb device selection) it will not start and the only way you would see the difference between not starting the kernel and driver not working would be to have a serial console connected (null-modem cable) if a live linux does not work i'd say check your network cable and switch

no, thats not the dmesg of the dsm system, you would need to mount the system partiton (raid1) if using a live linux but why a live linux? you did say you have 3617 installed on a ssd on the sata cdrom connector and the sata/sas on the perc h330 is not detected so you just boot up that 3617, ssh into it and copy the /var/log/dmesg to a share (like /volume1/data/) or to a usb you connect (if you can use winscp then this is ok too)

try this again but with a mismatch of vid/pid can you copy the /var/log/dmesg from the ssd disk (booted with at least one other disk attached to the lsi sas)

it does, if it boots and you find it in network its working, if the cpu is not up to it you would not find it in network if 918+ throws error 13 you can try with a wrong vid/pid, cant get worth then error 13 but as you cpu does not have gpu there is nothing to gain with 918+, more likely you will have driver problems with the lsi sas controller (-> 1st post, warning about disk hibernation and broken raids) 3617 seem to be the best choice but 3615 woul work too, as the h330 has a older lsi sas chip it should work with 3615 too in theory you could shutdown the r420, take its working usb and boot with it on the r430 before installing with it you should at least save the zImage and rd.gz because when installing the might be overwritten (when installing a newer version of dsm) that way you know its not the usb (as its working on r420) and just so see if disks are found it should be enough to just boot it and look with the browser on the systems ip address if its not booting at all or not found in network then check csm settings in bios and usb boot device (non-uefi to get csm mode booting, if you use the uefi usb boot device it will ignore csm mode as its just an option that gets chosen by selecting the right boot device)

dann schau dir mal die files /etc/VERSION und /etc.defaults/VERSION an, sollten bei den gleichen Inhalt haben, wenn nicht ist evtl. ein update auf eine neuere DSm version gelaufen die dann mit dem alten laoder nicht mehr gestartet ist (könnte man auch als loader defekt interpretieren obwohl das nicht booten da andere ursachen hat) in dem file steht zielich genau drin welche die lädst du dir runter https://archive.synology.com/download/Os/DSM und öffnest sie mit 7zip, da drin findest du u.a. zImage und rd.gz die müssen beide auf den loader und ersetzen die dateien die auf dem laoder sind wobei die rd.gz in ramdisk.lzma umbenannt wird dann exvtl. noch das Update zu der version über die web gui dazuinstallieren und dann sollte es hoffentlich wieder gehen abhängig von der hardware kannst du gleich auf 6.1 oder 6.2 updaten und den dazu passenden loader verwenden https://xpenology.com/forum/topic/7848-links-to-loaders/ wenn das board ahci bei sata kann dann sollte auch 6.1/6.2 gehen, mal mit "lspci -k" nachsehen grade 6.1 sollte da einfach sein da man nichts in sachen uefi/csm beachten muss https://xpenology.com/forum/topic/13333-tutorialreference-6x-loaders-and-platforms/ https://xpenology.com/forum/topic/7973-tutorial-installmigrate-dsm-52-to-61x-juns-loader/ poste doch mal was in den beiden VERSION Dateien steht und was "lspci -k" auswirft ansonsten kannst du auc die /var/log/dmesg ansehen um zu schauen was beim raid evtl. nicht geht (evtl. log hier hochladen) "cat /proc/mdstat" gibt die grob den zustand der raid sets aus

not in IT-mode? if you go into the controllers bios there can be IR mode or IT mode, if its IR (R like Raid) the it should be reflashed to IT mode you can look for guides on the internet (and even youtube) https://forums.servethehome.com/index.php?threads/flash-crossflash-dell-h330-raid-card-to-hba330-12gbps-hba-it-firmware.25498/ if you want to see disks in IR mode you would need to define a raid set in the controller, but thats not supported by native drivers and brings along other problems, dsm is made for single disks with software raid did the "wrong vid/pid" do anything, like being able to install?

we had a case lately where someone deliberately entered a wrong vid/pid after failed boot attempts, maybe try this as its easy to do https://xpenology.com/forum/topic/42899-usb-stick-brand-speed-issues/

is documented that 3615/17 loader 1.03b dont work with uefi https://xpenology.com/forum/topic/13333-tutorialreference-6x-loaders-and-platforms/ the little snag here is that if the 3615/17 loader is in GPT that makes uefi boot possible in the 1st place even if its not woeking, and if you dont select the non-uefi boot device the csm mode will nor be used the better solution might be to only offer then3615/17 in MBR version, MBR cant be used in uefi mode and CSM has to be used, kind of ride on rails that way, you you try a MBR loader with out CSM active you bios will tell you its not working this way and you will be steered to CSM mode if CSM is not active you usually dont see the non-uefi boot device and even when activating CSM in bios you usually need to save and reboot to see the non-uefi boot device in the selection list ("older" often have csm on by default but newer system come with csm disabled or even dont have csm anymore)

when modding the config file manually you can have 24 drives, but these is lost on bigger updates and needs to be re-done after the update the base config is 12 and the extension units are handled differently in dsm

größere updates die ein volles image der systempartition mitbringen, in der regel 200-300MB groß wenn du auf 6.2.3 bist wird es das erst mal nicht geben da 6.2.4 nicht mehr mit dem loader läuft und eher zu erwarten ist das es mit 7.0 weiter geht (ein loader der 6.2.4 kann wäre vermutlich netter abfall bei 7.0)

die default config hat 12 disks, damit sowas passiert müssten die disks mindestens oberhalb der 12 ports liegen, was eigentlich bei reinem onbord sata nicht passieren sollte da die immer am anfang liegen und oben drauf die zusätzlichen controller kommen man würde in /var/log/dmesg sehen wenn ein controller mehr sata ports blockt als er ports hat hat man bei älteren boards häufiger gesehen das die 4xsata hatten und zwei nicht genutzte ports trotzdem immer für eine p port "lücke" in den disk slots in der dsm gui gesorgt haben extrem fall ist bei mir ein asm1166 basierter 6 port controller der immer 32 ports blockt ach ja und der nachteil mal eben mehr ports als den default des laoders einstellen bedeutet bei updates verlust dieser config und das raid ist erst mal nicht funktionsfähig, je nach dem wie viele disks "fehlen" hat das folgen, wenn nachher weniger disks als für ein funktionierendes raid da sind startet das raid nicht und wenn man die config dann anpasst und er alles sieht dann startet das raid wieder ungünstiger kann es werden wenn so viele platte das sind das das raid startet, dann sind die nicht gefundenen redundanzplatten aus dem raid (weil es gestartet ist) und man muss ein rebuild der platten anstoßen (was 12-48h dauern kann) und in der zeit hat man keine redundanz ich empfehle immer erst nach einer anderen lösung zu suchen, in 1-2 jahren hast du vergessen das du die konfig geänderts hats und wie das geht und dann bricht panik aus wenn nach dem update plötzlich das raid futsch ist

problematic choice as it uses "Intel® I225-V 2.5Gb Ethernet" and that "igc" driver is only part of kernel 5.x and there is no standalone driver for older kernels (and itel stated they have no intentions to make one) - just dead weight and you need pcie nic card for dsm problem with nvme is that it can only be used as cache in the normal configuration, easiest way is to use a hypervisor and make them available as normal ssd in a vm but you can't have a dsm system with just two nvme's you would not be able to even install dsm as cache drives dont hold system and swap partitions just to use them as cache even needs patching and the cache will be "lost" on some updates so it needs to be disabled before updates https://xpenology.com/forum/topic/13342-nvme-cache-support/page/3/

130W TDP, hope you live in a cooler region i guess its loader 1.03b as 1.02b would only work for dsm 6.1 3617 is the better choice for lsi sas hba's as it comes from synology with newer drivers then 3615 3615/17 have a build in limit of 12 disks, if you dont disable the onboard sata the disks on the hba will be above the 12 disk limit (the 12 ports count even without a disk connected) and wont be used in dsm if you disable all the onboard sata or maybe keep max 4 ports (usually its per controller, disabling just one port of a controller usually does not work, its disabling or enabling the whole controller) on then your disks on the hba should work i'm not so sure you will see any increase in performance (that can be "seen") unless you use ssd's, normal hdd's hardly get more then 300MB/s (the cache of the disks can do more but its usually the disks that limits the speed) your 6G sata pors ore no better as they are only connected to pcie 2.0 as Intel X58 only supports pcie 2.0 https://www.asrock.com/mb/Intel/X58 Extreme6/index.asp - 2 x SATA3 6.0 Gb/s connectors by Marvell SE9128, support RAID (RAID 0, RAID 1 and JBOD), NCQ, AHCI and Hot Plug functions - 4 x SATA3 6.0 Gb/s connectors by Marvell SE9120, support NCQ, AHCI and Hot Plug functions (SATA3_6 connector is shared with eSATA3 port) https://www.marvell.com/products/system-solutions/sata-controllers/pcie-2-sata-controllers.html 88SE9128 (PCIe 2.0 x1 to 2 6Gb/s SATA ports) 88SE9120 (PCIe 2.0 x1 to 2 6Gb/s SATA ports) so it might be 2 x 88se9120 and 1 x 88se9128 with 3 pcie lanes and 500MByte/s per lane so every sata port has just 250MB/s, not better then the SATA II ports with 3G - basically there are all bad in the same way and not suited for ssd's you should check the disks performance before and after hba use, you can't control the hdd's cache when connected to the hba

if replacing a synology driver its usually a newer driver supporting more hardware, so synology branded hardware should always work (like the tehuti 10g nic driver, synology only supports one phy they use but the replacement driver supports all)

you are on 6.2.3 u3 (working) and want to update to ... 6.2.3 and it does not work? there seems to be something missing that thread is specific for the downgrade, for other problems please open your own thread in a appropriate section (and imho its no good having just one line from the log, it needs context and without dmesg it will be hard to see what disk got what number - and no information about the systems hardware or what type of dsm or loader was used ... i'm sure you know exactly about these things but keep in mind we don't and you need to give information when you expect to get help or advise)

on my systems on dont see a kernel crash when booting uefi with 3615/17 there is nothing on the serial console at least it work now within expectations, 3615/17 should work with amd

istalliere dsm auf eine virtuelle (kleine platte) ohne ein daten volume anzulegen, dann gibt es nur system und swap (2.4+2GB) auf der disk dann hänge die beiden anderen dazu (am besten be laufendem system (wenn proxmox das kann) dann bindet man die disks ein als sagen wir mal raid1, dabe werden sie intialisiert, was bedeutet das system und swap drauf kommen runterfahren und die hilfs disk entfernen, jetzt sollte der loader ie anderen zwei platten mit system (raid1 über alle disks) erkennen und booten, die fehlende nur system platte ignoriert dsm einfach, sind ja 2 im raid1 drin ist im wesentlichen das hier https://xpenology.com/forum/topic/12778-tutorial-how-to-downgrade-from-62-to-61-recovering-a-bricked-system/ nur das deine zwei platten leer sind

there are a lot of nice looking components that will things go south, you need to look very careful for specs and chips on cards like this one, hyper duo ? sounds like marvell, might only have one at best two pcie lanes and can only do pcie 2.0 i looked it up and its a 88SE9230 so its 2 lanes pcie 2.0 resulting in max 1000MB/s for the 4 drives, so not optimal the M.2 is pcie 3.0 x4 and there are M.2 to pcie 4x slot adapters with flex cable, so the m.2 can become a pcie 4x slot https://www.amazon.com/ADT-LINK-Extension-Express-Extender-R42SF/dp/B08DRDT47K (i bought two, the shorter one above works so far but i did not test much as its in my backup system with 10 disks and kind of crammed) the pcie 16 slot looks like as it has 16 lanes and is pcie 3.0 maybe a jmb585 based card that supports pcie 3.0 (even with just 2 lanes its way better then the old marvell cards with just pcie 2.0 read this thread about some options https://xpenology.com/forum/topic/35882-new-sataahci-cards-with-more-then-4-ports-and-no-sata-multiplexer/ i fully agree with @flyride ssd cache might not be worth the risk and cost, 16GB or 32GB ram do some full 10G speed buffering too might be better to have a ssd volume for special purpose like vm/docker using the 16x slot for a two or four m.2 ssd seems not a option here as the cheap adapters need bifurication to work and the manual does not mention that as flyride suggested try 10G nic first and see how it performs and then decide how to go further ssd cache makes no sense with 1G nic you won't see any difference your 8 disks should easily saturate that nic i also use a matx but one the has lower depth just 10mm more the m-itx ()but its wider and has 4 pcie slots), getting a new housing might be cheaper and easier the a m-itx with 10G onboard and enough sata options to have 12 or more ports (i know that as my system has 12 ports and 10G and i had m-itx before)

i think i did once and the synology help also shows it as possible with a additional disk https://www.synology.com/en-global/knowledgebase/DSM/help/DSM/StorageManager/storage_pool_change_raid_type