IG-88

thats a problem as 16 cores (8 real, 8 HT) can only be used with 3617, th other two are limited to 8 cores (kernel compile defaults from synology) to use nvme you need 918+ (you cpu is haswell so it should work) but you would need to disable HT or you loose performance when HT is used instead of real cores (i dont know how the kernel chooses but it its 4cores/4ht it would be significant) i'd suggest going with two normal sata ssd's and use write caching (only with two drives allowed) afaik ssd cache is only strong in multi user environment, m most home use cases its not that much of a performance boost as it only serves you data you already have in ssd cache faster, also to make good use of ssd/nvme you will need more then 2.5G (still much better then one), 10G for nvme and 5G for sata ssd or you will loose significant on performance edit: the only way to use nvme with 3617 might be esxi and then doing a rdm to the xpenology vm, but that way you would do all other VM's with esxi and xpenology would be degraded to be just a storgae system, also wehe using esxi it might be possible to just user 4 cores for xpenologgy and map the pcie nvme device into a 918+ vm? yes that sounds tempting

the link ended in a "not found" so i cant say anything about if its supported by the additional driver package if its 2.5GBit only it might imply a realtek 8125, witch is supported the micro atx board i found with this chipset would be HUNNANZHI X99-8M ~130$ Asrock FATAL1TY X99 M Killer ~400$ not sure why you go with that outdated hardware/cpu maybe a 8 core ryzen might be a better choice (as 8 core intel wil cost double) ? you did not say much about your base for choice so its hard to suggest anything else

i'm not so sure this does look ok together beeing cheep on such a important part? if its important then maybe a known server board brand like supermicro would fit the description? ??? the board seems to have just a 1Gbit nic, if it would be real the it would be a NBASE-T nic and the description did say 1GBit or is it any additiona nic you are planning?

/etc/sysconfig/network-scripts/ ifcfg-eth0 ifcfg-eth1 ... in these files set IPV6INIT=off

the patch doing the work is inside extra.lzma, etc/jun.patch you would need to unpack extra.lzma, make a diff against the original syno.info.conf from the *.pat file with a syno.info.conf patcht the way you want it (first apply juns patch, the you won 24 disk mod and the make a diff, that diff will replace the parts inside jun.patch - that it was i would do)

if you have shr then you can exchange four 2TB disks one after another (rebuild might take one day per disk) and after you are done you would have 20TB shr in 5 disks, four 6TB, one 2TB, as the 2TB disks are older you can keep two as spare in case the one 2TB disks in your shr fails better way would be the migration tool and transferring to a new or interim hardware, might take longer (1GBit nic?) but you get rid of the 5th drive, i never used Migration Assistant but i guess its exactly for this kind of operation

you did not read the faq ? even in default state it will use a existing shr raid, the limitation is that you cant create a new shr, to do that you just need to change the synoinfo.conf and that limit is gone thats because you did use 6.2.2 pat file, if you use 6.2.(0) you would have been able to use it 6.2.2 needs new drivers (inside the file extra.lzma) https://xpenology.com/forum/topic/21663-driver-extension-jun-103b104b-for-dsm622-for-3615xs-3617xs-918/

@wanye are you sure you need the few things you got from updating, you might just use the 2nd option in the loader and do a fresh install, your raid/data partitons should not be effected, this step would on do a fresh install of dsm on the system partition(s) it might be much less time consuming to redo the ip address and user/group, shares configuration, the plugins would be needing reinstalling new versions anyway even if ou could use the updated dsm you also need diferent drivers when using 6.2.2, the loader 1.03b has drivers that will fail in a lot of cases with 6.2.2 as the "original" drivers from jun a made for 6.2.0 and synology made some kernel config changes with 6.2.2 https://xpenology.com/forum/topic/21663-driver-extension-jun-103b104b-for-dsm622-for-3615xs-3617xs-918/

you might want to read this, its a good way of understanding the process for recovering a raid system https://xpenology.com/forum/topic/24525-help-save-my-55tb-shr1-or-mount-it-via-ubuntu/ a easy estimate about the potential loss of data will be to know the sequence number of the disks, to get something you can access again you would need to force one of the disks into the raid set - meaning loss of data, - and then rebuild the redundancy disk but before doing this you need to know what caused the problem and make sure it does not happen again, as in the case above seen it will be a fruitless effort trying to repair a unstable system in a "real" recovery where your date would be worth money you would clone all disks and work with copy's or image files on a proven stable hardware, professional would not try to repair on a system with unknown's this will tell you more about the state of the disks mdadm --detail /dev/md2 mdadm --examine /dev/sd[abcd]5 | egrep 'Event|/dev/sd' also check log files and check s.m.a.r.t. status of the disks, you need to know how it happened are you sure? i can tell you from personal experience that even a 2 disk redundancy is not enough and there can be also file system corruptions just to make sure - you need backup, not more redundancy btw. did you ever thought about possible "other" scenarios? theft, fire, ... in some cases it just needs one big usb disk to cover the most valuable data or maybe cloud storage like this https://www.backblaze.com/blog/synology-cloud-backup-guide/ https://www.backblaze.com/b2/partner-synology.html (i followed them over the years because of the storage pod designs and disk statistics - having data in the cloud is not my thing but might be a option for some people)

no, juns patch leaves the original of 12 and my extra.lzma only adds srivers i guess yes but your 24 drive cout need to be done the way it was done before, manually changing synoinfo.conf ??? what did you change? there are no parts in the patch about max drives and the disk layout (binary pattern as hex number) you would need to do a new diff file against the original and a completely patched one (juns path plus change for more drives)

your microserver gen8 is documented to work with the new driver package for 6.2.2 your 1st attempt would have been ok with a usb flash drive with 1.03b and the newer kernel files a the new extra.lzma the method of doing the usb 1st is to prevent people from (semi) bricking there system with the 6.2.2. updaten and then having trouble preparing the new usb if the usb works as 6.2.2 with the new drivers then there is not much that can go wrong

if its more important then using the latest dsm version you can try older loaders with dsm vsersions 6.0 and 6.1 also look for anything that says "wake on ..." and disable it you try to disconnect everything thats not needed like usb devices (you only need the boot drive), keyboard, mouse, monitor, ...

that means your problem is not the same, thats a good documented thing (from my perspective) a failed driver prevents s shutdown at all, not even a reboot your problem is that the shutdown is completed and after that the computer wakes up again within a few seconds and boots (that not that uncommon, i've even seen this behavior with windows 10 systems, not just with linux)

its about getting the acpi power button to work, if you have general shutdown problems like yours (you cross posted here) then this package can't solve it

ich habe mir ein desktop gehäse entsprechend umgebaut aber neben der arbeit die das macht ist es auch nicht hotswap und für 12 platten, außerdem sollte es in ein 40cm tiefes regal passen (gehäuse als ~35cm tief) und flach sein, so das ich wegen der vielen vorgaben auf hotswap verzichtet habe, leise war wichtiger (wegen wohnzimmer) aber bei der suche kann ich die caseking.de empfehlen, die haben eine gute suchhilfe mit filtern und auch vorgedämmte gehäuse wenn man weniger probleme mit der tiefe (es 40cm gehäustiefe sein können) und höhe hat gibt es da sicher eine auswahl (auch mini tower die man auf die seite legen kann) und mit einschüben für 5,25" bekommt man auch hotswap nachgerüstet (wobei ich da wegen der dichte der platten skeptisch bin und fast immer ein lüfter mit verbaut ist der sicher nicht leise sein wird, aber mit ausreiched tiefe im gehäuse kann man da evtl. mit einem eigenen 80-120er lüfter nacharbeiten) leise nas gehäuse sind aber eine zu kleine marknische als das man sie hier findet, evtl. findet man in china shops wie aliexpress oder ähnlichen etwas passendes? die meisten consumer landen in so einem fall bei qnap oder synology, selbst bauen ist nur für sehr wenige eine option

as its a script anyone who has this "unusual" case can just change the path in the script to volume2 or anything else, it just needs to be documented so people having trouble can check for this condition

if you can't use CSM and have a "newer" cpu (like gemini lake) you still can cross upgrade to 918+, the lsi controller should be working there too and as long as you dont need the raidf1 for for ssd only systems there is no loss loader 1.04b and 918+ works with uefi

also wenn ich den im internet suche dann finde ich einen 4Gb Fibre Channel HA, keine netzwerkkarte ich habe einen broadcom chip BCM57840, den gibts anter allen möglichen oem namen (ich habe eine von asus, ich wollte dualport rj45 und das wer etwas billiger als intel) außerdem habe ich auch zwei günstige tehuti karten, eine ältere die ootb mit dsm läuft ein eine mit einem neueren phy der aber mitlerweile vom zusätzlichen treiber packet auch erkannt wird so das die jetzt auch funktioniert (hatte ich eine weile in der windows kiste weil es nicht mit dsm ging) - imho sollten alle tehuti based kartem gehen, ich hatte alle phy treiber die im treiber benutzt wurden eingesammelt und mit compliliert aber auch die günstigen aquantia karten (z.b. von asus) sollten gehen aber auch die gebraucht recht günstigen mellanox karte soltlen gehen auf jeden fall alle die ICH habe (selbst testen kann) und brauche aber davon abgesehen schaue ich das ich möglichst die gängigen/preiswerten 10G karten im treiber packet unterstützen (sofern es passende treiber gibt und ich zeit habe), also wegen mir musst du jetzt kein Noctua nehmen, ich habe ein desktop gehäuse, da gibt es ein wenig grenzen nach oben und mein kompromiss war der Noctua NH-L9x65 (4 heatpipes) wenn du jetzt einen tower nimmst und der nicht grade im wohnzimmer steht hast du wesentlich mehr optionen, der L9i wäre zwar möglich aber doch eher unnötig begrenzt größere lüfter heißt häufig auch geringeren geräuschpegel und mehr heatpipes meist mehr wärmeableitung (zumindest werden 2 weniger schaffen als 4, da aber die fläche der cpu begrenzt ist wird es nicht unbegingt so viel besser bei 4->6 heatpipes?) bei 65W cpu ohne overclocking ist eine 4 heatpipe lösung sicher ausreichend aber mehr schadet in dem fall auch nicht, im "schlimmsten" fall kühlt das monster die cpu passiv weil alles gut abgeleitet iwrd und der luftstrom im gehäuse ausrecht (bei 8-12 platten braucht man ohnehin einen gewissen minimalen luftstrom um die platten bei laune zu halten - mach ich zumindest so und der weg den die luft nimmt ist bei nas anders als bei gaming systemen, ich habe mein desktop gehäuse dahingehend angepasst, die normalen lüfteröffnungen (seiten) sind zugeklebt und an der unteren vorderkannte habe ich zusätzliche löcher gebohrt

whats with your usb? did you take it from the old system to the new what loader and dsm version? loader 1.03b for 3615/17 does nor support uefi, you would need to activate CSM mode and choose the non uefi usb boot device https://xpenology.com/forum/topic/13333-tutorialreference-6x-loaders-and-platforms/

well even after asking 2 times you have not documented what you have done in you synoinfo.conf good luck with your problem, i'm out

kommt darauf an was du an merkmalen so haben musst und für wie viele paltten wenn es in richtung desktop gehen soll, viele platten rein sollen und du nicht basten willst vieleicht das https://streacom.com/products/f12c-chassis/

J1900 seems to be the same as J1800 with 4 cores instead of 2 https://ark.intel.com/content/www/us/en/ark/products/78867/intel-celeron-processor-j1900-2m-cache-up-to-2-42-ghz.html https://ark.intel.com/content/www/us/en/ark/products/78866/intel-celeron-processor-j1800-1m-cache-up-to-2-58-ghz.html so syno might be the one to try, if you use the recovery type you can be sure that nothing i915 related will interfere and network should work (if you did all steps right) if recovery does not work for you its very likely that you missed a step and you would need to describe in detail what you did just to see it in network you can try just without changing the grub.cfg, the most important vid/pid matters when you install the dsm file, just to find it in network you can use jun's grub.cfg from loader 1.04b, it you find it in network with synology assistant then you can start changing grub.cfg (not messing around with grub.cfg might spare some time trying out things)

no you login with your admin account and then do "sudo su" to have root access if you just change the one in etc.defaults and reboot the other in /etc is changed automatically, so in a way you only need to change one file (if you reboot after changing it) and the good part is the way the scheme about the bit mask you need to set for the three (hex) values stiil it does work for me and a lot of other people, so its clear you made a mistake somewhere and if you want to know what ist is you need to communicate and explain what you have done

r8152.ko is already part of my extra.lzma (v2.12.0, 2019/04/29) just follow the instructions here https://xpenology.com/forum/topic/21663-driver-extension-jun-103b104b-for-dsm622-for-3615xs-3617xs-918/

what did you change exatly if you write the things you've done we migth spot the problem it should work with dsm 6.1 (or dsm 6.0) the same way i usually suggest this video https://www.youtube.com/watch?v=2PSGAZy7LVQ