Leaderboard

'Edit' is not available anymore - so I can't update original post.

you must restart pvedaemon and pveproxy - post #178

Und noch ein Addendum: Ich habe mal vor Jahren im Synology Forum herausgefunden wie man den Raid Rebuild/Build beschleunigen kann. Das ist mitunter hilfreich, es gibt aber auch Einschränkungen. Synology gibt dem Raid Rebuild nicht die höchsten, ich nenne es mal: "Task Prioritäten". Der Grund ist das man das System während des Rebuilds noch nutzen können soll mit guten Netzwerkgeschwindigkeiten. Kann man darauf verzichten geht der Rebuild deutlich schneller wenn man folgende Befehle als superuser eingibt: echo 32768 >/sys/block/md3/md/stripe_cache_size echo 200000 > /proc/sys/dev/raid/speed_limit_min echo 330000 > /proc/sys/dev/raid/speed_limit_max Ich hatte davor ca. 77MB/s schreibend. Nach dem Befehl nur noch 13MB. Die Weboberfläche reagiert weiterhin wie gewohnt aber der Rebuild ist deutlich schneller. Gruss, MPC561 PS: Es gab damals auch Anmerkungen das das nicht bei jedem System funktioniert weil nach ein paar Sekunden die Werte wieder auf die ursprünglichen zurückgesetzt werden. Konnte ich aber niemals bemerken bei mir hats mit 3 echen DS immer funktioniert. Mit cat /proc/sys/dev/raid/speed_limit_min cat /proc/sys/dev/raid/speed_limit_max cat /sys/block/md3/md/stripe_cache_size kann man sich die aktuellen gesetzen Werte anzeigen lassen. Aber keine Angst, kaputt machen kann man damit nix.

Hi @stasheck I finally managed to install Xpenology after having spent hours ! Even if everything is working now, I would not be able to tell you exactly what I really did well to make it work !!! (given that I discovered I had a faulty motherboard/bios not really able to deal with DSMAR table and so I had to use the unsafe_interrupt method within proxmox, and I am not able to tell how it might have affected my xpenology installation...) I hope you will get some more precise help form experts, but I can provide some partial outputs : 1/ I never managed to install 918+ nor loader 1.04, I was getting a fatal error during installation (I do not remember precisely error details) 2/ ds3617 xs was working with 6.2.1 but I was unable to update to 24922.pat or higher, it was also crashing 3/ I so use ds3615xs with 1.03b. I firstly installed 6.2.1 then update to 6.2.2 (I kept 1.03b) 4/ Yes I use e1000e nic drivers, implemented following the @blessendor method 5/ I did not managed neither to use virtual USB key for loading, I had to use sata disk. And the boot disk does appear in HDD manager but normally I depends of your choice in the loader boot menu (baremetal or VMware options) 6/ I firstly managed to get xpenology working and updated before any device passthrough I was finally able to passthrough 1 PCIe sata controleur card with 5 sata disks and USB + 3 other sata disks directly I hope this partial feedback will give you some clues. And do not give up, it should finally work fine also ! ... even if I am quite anxious about the moment the next update will show up !!! Good luck

You forgot the „magic crystal ball“ which everyone one of us should have 😉 (no offense to the TS).

Hi Guys! since i'm the starter of this topic i wanted to tell you what i've done i learned that my previously used sata controllers mentioned above didnt worked and that i have to use HBA controller from a known brand. so i searched one and finally bought a 8port "HP H220 IT Mode LSI 9205-8I PCI-e 3.0HBA" controller from HP. the one i bought from ebay was this one: https://www.ebay.com/c/22028264506 it came with all the needed cables. i just built it in, plugged my drives in it and all my drives where recognized on my mainboard i'm using all 8 ports with 8 drives. on the controller i tested it with all 8 slots and they are all working, now i can use 16 drives simultaneously 😮

It was the first reboot after starting Upgrade DSM - so my version was 6.2.2-24922 - but with not yet finished upgrade. There were two problems: 1) boot from USB img failed - fixed by replacing ich9-usb-ehci1 to nec-usb-xhci (found examples on the proxmox wiki page), 2) next I was must remove 'scsihw: megasas' string (I have 6 SATA and 2 SCSI disks attached to this VM, so I played with different SCSI controller model in the past on first-time setup). Solved.

After successful testing of the @hellbondz's method, I found that my second network interface in the DSM is inaccessible from my LAN (I use multi-homed network). This was happened because: Without this script PVE will add both netdev into default bridge - vmbr0 (a public bridge in my case). I patched two PVE scripts to be able add network interface with Proxmox GUI and staying out from edit config manually (what somebody can't do it correctly). So my method of using Network Driver Intel e1000e with Proxmox, which required to be able upgrade to the the latest DSM version (current up to DSM 6.2.2-24922 Update 2) and keep network worked: 1) Make DSM VM backup as usual! 2) Make PVE scripts backup: cp /usr/share/perl5/PVE/QemuServer.pm /usr/share/perl5/PVE/QemuServer.pm.default cp /usr/share/pve-manager/js/pvemanagerlib.js /usr/share/pve-manager/js/pvemanagerlib.js.default 3.1) Open /usr/share/perl5/PVE/QemuServer.pm find needed block with keyword 'e1000 (or 'vmxnet3' - less results than for '1000') replace: my $nic_model_list = ['rtl8139', 'ne2k_pci', 'e1000', 'pcnet', 'virtio', 'ne2k_isa', 'i82551', 'i82557b', 'i82559er', 'vmxnet3', 'e1000-82540em', 'e1000-82544gc', 'e1000-82545em']; to (add 'e1000e'): my $nic_model_list = ['rtl8139', 'ne2k_pci', 'e1000', 'e1000e', 'pcnet', 'virtio', 'ne2k_isa', 'i82551', 'i82557b', 'i82559er', 'vmxnet3', 'e1000-82540em', 'e1000-82544gc', 'e1000-82545em']; 3.2) Open/find/replace /usr/share/pve-manager/js/pvemanagerlib.js if ((match_res = p.match(/^(ne2k_pci|e1000|e1000-82540em|e1000-82544gc|e1000-82545em|vmxnet3|rtl8139|pcnet|virtio|ne2k_isa|i82551|i82557b|i82559er)(=([0-9a-f]{2}(:[0-9a-f]{2}){5}))?$/i)) !== null) { res.model = match_res[1].toLowerCase(); to: if ((match_res = p.match(/^(ne2k_pci|e1000|e1000e|e1000-82540em|e1000-82544gc|e1000-82545em|vmxnet3|rtl8139|pcnet|virtio|ne2k_isa|i82551|i82557b|i82559er)(=([0-9a-f]{2}(:[0-9a-f]{2}){5}))?$/i)) !== null) { res.model = match_res[1].toLowerCase(); next replace block: Ext.define('PVE.form.NetworkCardSelector', { extend: 'Proxmox.form.KVComboBox', alias: 'widget.pveNetworkCardSelector', comboItems: [ ['e1000', 'Intel E1000'], ['virtio', 'VirtIO (' + gettext('paravirtualized') + ')'], ['rtl8139', 'Realtek RTL8139'], ['vmxnet3', 'VMware vmxnet3'] ] }); to: Ext.define('PVE.form.NetworkCardSelector', { extend: 'Proxmox.form.KVComboBox', alias: 'widget.pveNetworkCardSelector', comboItems: [ ['e1000', 'Intel E1000'], ['e1000e', 'Intel e1000e'], ['virtio', 'VirtIO (' + gettext('paravirtualized') + ')'], ['rtl8139', 'Realtek RTL8139'], ['vmxnet3', 'VMware vmxnet3'] ] }); 4) Run 'service pveproxy restart' and reload PVE admin interface - you will be able to change the Network driver of your DSM VM from e1000 to e1000e. 5) Save copies of edited scripts to somewhere else to be able restore/compare scripts versions after Proxmox upgrades. Again - make backups before run any test from my tutorial. Good luck

My NAS (see in my signature) cost me ~200$, including all parts. Though it is true that I bought the PSU and case used. The HW itself is a lot stronger than the Microserver, and possibly outdoes the T20 (you did not say which config it was exactly). You can go a bit higher and get an i3-7100, or a 6200, and still be under 300$ in total.