IG-88

as long as you have 12 disks and and they are connected to the first 12 ports you should not have any problems, its only a thing if you use a port above the count of 12 as this is the default for jun's loader (more or less by accident as 3615/3617 come with 12 from synology and jun refitted this "12" into his patch as he created 916+/918+ to have it as much like the "standard" used by most users until this point) if you want a "smooth" update with more then 12 ports i see 2 ways (both not tested as i dont need it for my system) 1. install the update and when it reboots after the update (your 24 disk config was replaced by the default config from synology) do not boot the 1st time with your normal usb flash drive, use one with a live/recovery linux, after booting this mount the system partition (mdadm raid1 over all disks, this steps are part of the downgrade version howto), change synoinfo.conf's how you need it to be, reboot with normal dsm loader flash drive 2. (more difficult but universal if its done), juns loader patches the orgininal synoinfo.conf it it finds the "old" (aka original synology parts) synoinfo.conf entrys and if there will change it to what you usualy see with jun's loader, you would have to extract extra.lzma, rewrite the patch file inside extra.lzma to change the values in question to fit 24 drives and repack extra.lzma in case of 916+/918+ its changed from 4 to 12, so should be easy to to as you can replace jun's values by your own, in case of 3615/3617 its much more complicated as the synology default is already 12 and jun's patch does not contain parts in that direction so you would have to extend the patch by creating new parts, needs more skills but possible if done correct it only needs to be redone when creating a new flash drive or put in another extra.lzma (like for newer drivers)

well also thanks to @jensmander as he has seen i was on the wrong way by assuming latest version to be the latest 6.1.x i even go further on my system, just as shown plus no internet access for the system (blocked in firewall), updates are only installed manually after checking not to make the system unaccessible (imho bricking is not the right word as all data are still there and can "easily" accessed or recovered - brick would be like its just a peace of dead rock without hope of recovery)

this will happen if you installed a newer dsm version with this usb flash drive, the update process overwrites the boot (kernel) files and contains the last version installed without downgrade option, replace content with 1.02b original image file (dont forget vid/pid change) and try 6.1.7 again

you cant use a new major release of dsm without a new bootloader 1.01 is for dsm 6.0, jun's 1.02 is for dsm 6.1, jun's 1.03 is for dsm 6.2 for now 1.03 is only alpha (testing) release and only for the 918+ (will replace the 916+ from 1.02 i guess) dsm 6.1 has bootloaders for 3615/3617/916+ whatever you use of this three, if you let dsm do a update to 6.1x to 6.2 it will fail because the bootloader (aka the hack integrated to mage dsm run on normal non synology hardware) does not match, no 1.03 for this versions for now my assumption was latest dsm = latest dsm 6.1 as it should be common knowledge when using this hacked dsm stuff you can't update to different major version there ist also a announcement area (in you scroll your browser to the top most ares) clearly to see when you use the website DSM 6.2-23739 - WARNING 05/23/2018 This is a MAJOR update of DSM. DO NOT UPDATE TO DSM 6.2 with Jun's loader 1.02b or earlier. Your box will be bricked. You have been warned. https://www.synology.com/en-global/releaseNote/DS3615xs

yes. default is 12, i tested with one disk on port 4 and another on port 12 maxdisk is 12 anf internal disk 0xff (as in 916+) but it shows 12 disk and seems to work

there is more then one? error 13 after updating? you used 1.02b and DSM_DS3615xs_15284.pat all was fine and then ... you could run a ram check from a bootable linux media, check s.m.a.r.t. of disks anything in the log?

i did send you a link to the test version with ixgbe v5.3.7, lets see what happens

jun used ixgbe driver version 4.2.1-k (from kernel 4.4.59?,) in his 1.03a2, in original dsm 6.1 3615 there was version 4.4.6, my latest extra.lzma for 1.02b (3615 working for you) used 5.3.3 so i might have to replace that too by a newer version with the extra.lzma for 1.03 i remember i had someone with a 10G intel driver last year and used newer source from intel so i will check out intel for the latest version of intel nic drivers i hoped to get rid of all extra stuff when using kernel 4.4.x but looks like i will have to use external sources from the start (already had to for mpt2sas/mpt3sas) and the number is increasing

what will you gain by using 3617/916+ ? as the cpu does not have intel quick sync video you won't have hardware transcoding if you aim for 916+ and want to play a little why not trying loader 1.03a2 and 918+ dsm 6.2? i got it running on a C2550 (installed the disk on a different system as the install process always gave error 13 and vid/pid and files where ok) but that was just for testing, still running 3615 and its fine imho there i not much you can do beside bios settings, the kernel and system comes from synology and when its not coming up and its not missing drivers then there is not much you can do, compiling a new kernel is no option

if you need access now you can install open media vault to a empty disk, boot it an access the dsm drives, omv should be able to use the raid/lvm if you don't care about settings and plugins you can do a migration to different model like 3615 -> 3617, that way the system partition has to be completely rewritten, after that you can leave it or migrate back to 3615, your data should not be affected by this as its still just a migration also you should use loader 1.02b not the older 1.02a

sorry did not wanted to be rude, giving as much information about the config might help to figure out for others if they might have a similar configuration and how likely it is that it may work, we are trying to document where its working and where not and to find out why (so far 4th gen intel core seems to be minimum for 918+) i don't know why jun did this, it was the same on 916+ so i guess he had a reason but a i cant see it, i'd tend to change internalportcfg for 12 disks as used by synology in 3615/3617 on a 12 port system i had two disks for testing on port 4 and 12 and the disk on port 12 was working in the same way as the disk on port 4, so its not a problem in general not having a user root is pretty much normal on linux today, has nothing to do with synology but in general i'd say complains are for synology, suggestions for xpenology instead of using root you can modify the /etc/sudoers change %administrators ALL=(ALL) ALL into %administrators ALL=NOPASSWD: ALL and you should be able to use your admin account with winscp compared to 916+ its "only" the newer hardware transcoding support (kaby/coffee lake)? hopefully more/better drivers then with kernel 3.10.x on 3615/3617 (at least i got mpt2sas working with 918+, i did not get this working on 916+)

the detailed list with the collected feedback is this (for 918+ anything older then Haswell, intel core 4th gen, is not working)

you don't write anything about your hardware, at least cpu should be mentioned because its kind of common for 3615/3617 and a good ammount of people user more then the 4 disk default, it sprare's tjem tinkering inf if you use less it does not hurt you can activate root (as user), use the search, beside this whats wrong with sudo or just once sudo su ??? if you dont want 6.2 use loader 1.02 and dsm 6.1(.7)

does it show up in network after loading the realtek driver?

... it was a answer to BigBro's question not to yours atm i'm more interested in getting the drivers for 918+ done and sum up reports on working / non working 918+ attempts in the developer thread then answering individual problems with hardware accel. (i'm not using myself, my nas has a C2550 and the backup system is older then this) i might come back to this when the drivers are running and i have tested them on a skylake desktop hardware (that might support hardware accel.) there are other people already unsing hardware accel. since 916+, so mayby someone else might way in here

we already have Core i7-7700, 4 cores, 3.6 GHz base, 4.2 GHz burst, HD Graphics 630, Kaby Lake, Q1/2017 as working so it should work from the cpu side, we already had reports for i217 not to work, maybe i219 is causing trouble too try to disable/remove all adaters, just leave one active and try with just one and if it fails try with another

the nic might be a problem in barematal as jun does no provide all firmware files in his extra.lzma but in vm it not relevant, if you use vmxnet then there is a driver in juns extra.lzma the processor might be the problem if it E5520 or simmilar the its Nehalem EP Q1/2009 and thats way older then the 4th gen intel core desktop cpu we have found as working min, yet (see my list on page 5), so not likely to work (on the other hand no gain with 918+ image, dsm 6.1 works stable so why using 6.2, nothing to win a lot to loose)

hmm, the fact that it coems up without a disc idicates the kernel and realtek network driver are working,, that board has 2x intel onboard (ahci) and 2 ports ASMedia ASM1061, also ahci with kernel 4.x (https://cateee.net/lkddb/web-lkddb/SATA_AHCI.html) maybe you can try disabling the asmedia in bios for testing? or try installing 6.2 on a different system and after installing use flash drive and hdd on th Q1900

i've see that with my Avoton C2550 cpu too, everything was right (usb vid/pid) but install failed, so i did tried the same usb flash drive with a skylake system and after insalling i connected the disk and usb flash drive to to the avoton system and it came up nicely your cpu is older then the avoton, its Cedarview Q3/2011, your nic is e1000 driver and thats inside juns loader already, so maybe it will work if you install it on a different system and then use the disk on the thecus, as this is pretty early i would not advise to do it on a system you need for daily use, it could be that if a update comes up that it will not install cleanly, we would need more time to see waht happens with major updates like 6.2.(0)->6.2.1

its about the original kernel used in the loader, its compiled with settings by synology to have a good support/performance with the used 918+ hardware (J3455), not to have a wide range of other hardware supported as jun's approach is to use the kernel and dsm unmodded and circumvent the protection (check for certain hardware presence, signed drivers) we will have to live with the limits of the original kernel (among others no explicit amd cpu support, no hyper-v support), so the list about supported cpu's is for finding out where the limits are (drivers in most cases can be fixed, missing kernel option that need to be compiled into the kernel can't) i started the cpu list because jun stated in his comments to 1.03a2 that its unknown for now what will be supported and also on what happened with the 1.02 916+ image, nobody could really say in what hardware the quick sync video transcoding support would work, we knew min. 4th gen intel core to boot from jun an no one documented about the transcoding as the 918+ has a newer kernel 4.x (and hopefully will support more hardware then the old kernel) i'm kind of "count me in" for the 918+ if you look into the 3615/3617 dsm 6.2 images you will see they still use the old 3.x kernel so no improvements are to expect in this area, kind of dead end as long as we find somone fixing the sata_ * and pata_* driver issue in the synology source (maybe we will have the same in kernel 4.x, we will soon find out)

BCM5720 would be tg3 driver and thats part of juns extra.lzma. so the nic should work in baremetal BUT you cpu is sandy bridge and that did not work with desktop processors on 918+ image so i'd expect it to fail because of the cpu (see my cpu list on page 5)

please be more specific about what chip the nic is, at least the i217 seems to be a problem if you cant check about the chip please give the product name of the nic or the system board (if its onboard)

according to this https://en.wikipedia.org/wiki/List_of_Intel_Celeron_microprocessors#"Bay_Trail-D"_(22_nm) it seems to be on the level of ivy bridge so maybe not, but my guess for my C2550 would also have been no but its working so as long as we have no feedback on bay trail i would not write it off you can check on the list on page 5 of this thread or maybe in the then existing driver thread for 6.2 so see if we know more about your cpu already by then

on 1.02/DSM 6.1 i never replaced the e1000e driver as there where no problems reported (afair) and it was coming directly from synology the data from intel for e1000e (latest driver i looked up in 2017) are e1000e: Intel PCIe PRO/1000 82563/82566/82567/82571/82572/82573/82574/82577/82578/82583/Gigabit CT Desktop Adapter/PRO/1000 PT/PF/I217-LM/V/I218-V/LM/I219 LM/V so i217 should work at least with the latest e1000e driver (when used) i do have a Gigabit CT Desktop Adapter (chip on it is 82574) so i can test the driver but afair the adapter was working with the driver in 1.02b, but when using a recent driver from intel source i will be able to test if the driver is working i will have a look at the intel driver when i'm doing the 1.03 stuff this weekend, i guess i already compiled the driver with the e1000 driver and did not include it in the extra.lzma as there was nothing to fix/gain (don't try to repair thing that aren't broken) i'm pretty sure its the nic, you kind of confirmed that the i217 is not working i'm sure its not vt-d, my C2550 i tested does not have this, if you want to make guesses look for differences in gen3 vs. gen4 intel core cpu's, sandy and ivy bridge cpu's are not working with 918+ and if you add the C2550 to that instruction comparison ...

strange, the driver for the nic would be e1000e and thats inside the image and the platform haswell works for others maybe we can come back to this later when we have a driver package and have divers compiled from intel source, maybe a driver problem and not th cpu