Jump to content
XPEnology Community

unmesh

Member
 View Profile  See their activity

  • Posts

    167

  • Joined

  • Last visited

 Content Type 

Posts posted by unmesh

    Upgrading N54L to 10Gb NIC

    in AMD

    Posted

    I got the Mellanox NIC installed and adjusted grub.cfg with its MAC address but am waiting for my DAC cable and fiber modules.

     

    Will I see the relevant SFP information in the GUI or do I need to ssh into a shell and use the command line?

     

    Thanks.

    Upgrading N54L to 10Gb NIC

    in AMD

    Posted · Edited by unmesh
    Clarifications

    I've ordered the 10G gear (Mellanox ConnectX-3 and DAC cables) and, in the interim, re-enabled the built in NIC in BIOS and confirmed that shows up in the Web GUI. Also, it enumerated after the Intel ports in the added NIC which was helpful.

     

    Now waiting for the delivery!

    Upgrading N54L to 10Gb NIC

    in AMD

    Posted

    Fortunately, I am on 6.2.3!

     

    I have a managed switch with a couple of SFP+ ports and thought I'd try my luck with SFP+ NICs and DAC cables though now I'm wondering if I should first try link aggregation over a pair of gigabit connections since I don't have a high performance array. (Both servers have multiport NICs)

     

    I need to check whether my switch supports LACP or merely static LAG and what the support situation is on ESXi 7 and on DSM 6.2.3 on baremetal.

     

    Decisions, decisions!

     

    Thanks for the guidance.

    Upgrading N54L to 10Gb NIC

    in AMD

    Posted

    I'm currently running with a HP NC360T Intel-based Gigabit NIC and would like to upgrade to a Mellanox ConnectX-2 or -3 NIC. From what I've read so far, it would appear to be a best practice to leave a Gigabit port in the system. Also, I've seen a posting elsewhere that the iperf3 throughput tops out at 3+Gbps which is fine for me.

     

    Unless I've misunderstood that, is it possible to have the currently unused onboard Broadcom based NIC play that role given that there is only one PCIe slot available for a x8 card? And can I go directly from my current config to one where the Intel-based NIC is swapped with a Mellanox-based one?

     

    Also, how would I edit the MAC address(es) in the bootloader USB as I transition from NC360T NICs -> onboard NIC -> Mellanox NIC?

     

    Thanks in advance.

     

    P.S.: Once I'm successful with this, I hope to upgrade my Lenovo TS140 to a 10G NIC too though that involves ESXi 7

    Tutorial: Install DSM 6.2 on ESXi 6.7

    in Tutorials and Guides

    Posted

    On 12/21/2020 at 6:33 AM, luchuma said:

     

    MAC of phisical server cannot be duplicated

    You can use generator or let esxi to generate mac.

    I was thinking of increasing the real MAC address and putting it into both the grub.cfg and into the VM config before booting the VM. Otherwise I have to let ESXi create the MAC addresses (does it do it before booting the VM?), then edit grub.cfg and create the boot image.

    Tutorial: Install DSM 6.2 on ESXi 6.7

    in Tutorials and Guides

    Posted

    Hi guys,

     

    It's been a while since I did an install on ESXi and I've forgotten what to do for MAC addresses.

     

    Do I let ESXi assign a MAC address to the network adapter and then put that into grub.cfg? Or do I concoct a MAC address from say the MAC address of the server and put that into both the virtual machine config and grub.cfg?

     

    Thanks

    Upgrading hard drive(s)

    in General Post-Installation Questions/Discussions (non-hardware specific)

    Posted

    I'm lookin for advice on best practices for replacing hard drives with larger ones.

     

    Some background. My first foray into Xpenology was with the N54L when I threw in some spare hard drives in a JBOD configuration. Subsequently, I've spooled up a couple of Xpenology VMs each of which uses a large capacity RDMed hard drive and the N54L is used as a backup for one of the VMs.

     

    I would now like to replace the smattering of small drives in the N54L with a single large hard drive and was planning on doing the following. Add the new hard drive and let Xpenology create the partitions and copy the relevant software. Create a new volume and copy all the files from the old volume to the new volume. Delete the network share that points to the old volume and map it to the new volume. Delete the old volume and remove the appropriate hard drives.

     

    I'm expecting this to save me creating a new USB and doing a new installation. Am I missing something?

     

    Is there a better way to do this upgrade? Should I choose SHR to make future expansion easier?

     

    Thanks

    Recovering a DSM 6.2.2 ESXi VM installation

    in DSM 6.x

    Posted · Edited by unmesh
    Clarifications

    I accidentally deleted the folder containing the second of two DSM VMs on ESXI 6.7U3 and am hoping I can recover the installation because it was a simple one as these things go.

     

    I have a copy of the customized v1.03b DS3615x 6.2 synoboot file which I was using along with a single RDM'ed hard drive for the data and think I can recreate the VM settings from memory. However, I updated DSM from the GUI one or more times since the original installation (I'm guessing to 6.2.2 24922 Update 6 rather than U4 in my .sig) and vaguely remember reading that synoboot gets modified when that happens.

     

    The hard drive in question shows 3 partitions in ESXi, two of about 2GB each and a 3rd that takes up the rest of the drive.

     

    Before I fire up the VM, any pointers on how to proceed without losing the data on the hard drive?

     

    For the future, is there a best practice around creating copies of the synoboot file on DSM updates?

     

    Thanks

    ESXi 6.7U3 not shutting down from DSM

    in DSM 6.x

    Posted

    When initiated from the DSM GUI, the OS appears to shut down but not all the way as shown by the Vmware GUI. I set up a serial-over-LAN console and got the follwing output

      

    [  357.190206] init: synonetd main process (5393) killed by TERM signal
[  357.192923] init: ddnsd main process (9158) terminated with status 1
[  357.200603] init: synostoraged main process (9417) terminated with status 15
[  357.218126] init: hotplugd main process (9940) killed by TERM signal
[  357.233112] init: smbd main process (10212) killed by TERM signal
System is going to poweroff.
[  361.195232] init: synoscheduler-vmtouch main process (11420) killed by TERM signal
[  361.290355] init: Disconnected from D-Bus system bus
[  364.414348] iSCSI:extent_pool.c:780:ep_exit syno_extent_pool successfully finalized
[  364.416200] iSCSI:target_core_rodsp_server.c:844:rodsp_server_exit RODSP server stopped.
[  364.479316] EXT4-fs (md2): re-mounted. Opts: (null)
[  364.601074] init: skip respawn syslog-ng during shutdown
[  365.534343] md2: detected capacity change from 29421993984 to 0
[  365.535315] md: md2: set sdb3 to auto_remap [0]
[  365.536008] md: md2 stopped.
[  365.536456] md: unbind<sdb3>
[  365.536974] md: export_rdev(sdb3)
[  366.678609] md: md0 in immediate safe mode
[  366.679304] md: md1 in immediate safe mode
[  366.681192] init: cgmanager main process (4717) killed by KILL signal
[  366.682227] init: skip respawn cgmanager during shutdown
[  366.683194] init: dhcp-client (eth0) main process (7474) killed by KILL signal
[  366.684645] init: tty main process (10168) killed by KILL signal
[  366.685720] init: skip respawn tty during shutdown
[  366.688852] init: Failed to spawn dhcp-client (eth0) post-stop process: unable to connect to CGManager: Unknown error 196609
[  366.727126] EXT4-fs (md0): re-mounted. Opts: (null)
[  367.767601] sd 1:0:0:0: [sdb] Stopping disk
[  367.768376] sd 0:0:0:0: [sdm] Stopping disk
[  367.769306] e1000e: EEE TX LPI TIMER: 00000000
[  367.805455] e1000e 0000:03:00.0: Refused to change power state, currently in D0

    Any thoughts on what might be the cause?

     

    Running Jun's 1.03b bootloader and the latest DS3615xs DSM with a virtual hard drive

    ESXi 6.7U3 and multiple VNICs

    in DSM 6.x

    Posted · Edited by unmesh
    Fixed typos

    I added a second virtual NIC to a test DSM6.2 installation that was connected to the same vswitch as the first one, changed the grub.cfg to add a second MAC address, created new .vmdk files for the bootloader, rebooted and DSM showed the second NIC and an IP address for it obtained through DHCP.

     

    I then created a new vswitch connected to a new physical NIC and moved the second vNIC to this vswitch. This configuration will not pick up an IP address from the DHCP server even though the GUI "console" shows the relevant port icons as green.

     

    Do I need to change something with respect to the MAC addresses in the bootloader? The VM is configured to let ESXi assign them.

     

    Thanks.

     

    Added: I spooled up an Ubuntu VM with a single vNIC connected to the second (newer) vswitch and the network works fine.

    Accidental DSM 6.1.7-15284 to 6.2.2-24922 on ESXI 6.7U3

    in DSM 6.x

    Posted

    I've had a DSM 6.1.7-15284 VM running on Jun's 1.02b loader for a while and a DSM 6.2.2-24922 running for a shorter while on Jun's 1.03b loader. Each VM uses a different RDM hard disk.

     

    Earlier today, I accidentally hit update on the GUI for the first one when it offered to upgrade to 6.2.2-24922 and it (obviously!) won't boot correctly. So I replaced the loader in the VM with a suitably configured 1.03b, chose the ESXi option in the console, but it still won't boot DSM.

     

    I have the serial port on the VM configured to map to a network port which displays "mount failed" and nothing else

     

    Is it possible to recover the installation?

     

    Thanks.

    Using Multiple NICs on ESXi 6.7

    in DSM 6.x

    Posted · Edited by unmesh

    I haven't done stress testing but it absolutely splits traffic across both GE NICs as reported by both the Windows 10 desktop and the baremetal DSM. I suppose I could time the transfer with a watch too and report back to this thread :-)

     

    After some experimentation, I managed to reproduce the results under ESXi. I created a new vswitch, attached an unused GE port as an uplink, created a new port group that used this vswitch and added the port group as a new network interface to the DSM VM. I then configured SMB to use DSM3 through the GUI and by making an edit to smb.conf and was off and running.

     

    One thing that had concerned me was how Windows would know which two IP addresses to use for the server. It turns out that my network attached drive from the single port days was enough to make the connection.

     

    I had considered using 10G but put it off because I wanted to do multipoint and wasn't ready to invest in a 10G switch and a bunch of 10G NICs. I discovered that Multiport GE NICs are very cheap on Ebay; my i340-T4 was only $15!

     

    I will disconnect the cables to the second port everywhere for now and hope that gets rid of any potential instability or data corruption issues.

    Using Multiple NICs on ESXi 6.7

    in DSM 6.x

    Posted · Edited by unmesh

    10 hours ago, Olegin said:

    Show your XPEN VM configuration.

     

    Not able to access my server at the moment but the VM configuration has two VCPUs, 2GB of memory, a hard disk for the bootloader on an IDE controller, another for a RDM'ed hard drive on a SATA controller and a E1000 network adapter connected to vswitch0.

     

    The Lenovo TS140 came with a i217LM on the motherboard which shows as vmnic0 configured as the uplink on vswitch0

     

    The 4 new ports are showing up as vmnic1 through vmnic4 but are not configured to anything

     

    Since my opening post, I did try adding one of the new ports as an additional uplink to vswitch0 but that did nothing for file transfer performance.

     

    Using Multiple NICs on ESXi 6.7

    in DSM 6.x

    Posted · Edited by unmesh

    Having had good luck with using multiple NICs to increase file transfer speeds between a Windows 10 desktop and a baremetal Xpenology install, I'd like to do the same with a Xpenology-on-ESXi-6.7 system. To this end, I installed a quad-port Intel Gigabit NIC alongside the built-in i217LM and VMware sees the 4 new physical NICs. I'm at a complete loss for how to get an additional NIC mapped over to the Xpenology VM. I don't want to do Teaming but rather use SMB 3.0 Multichannel end-to-end so that the network does not have to be configured. I will leave the other three Gigabit ports unconnected for now.

     

    Should I add a physical NIC to the existing vswitch as a additional uplink or will the E1000 driver only allow 1Gbps of throughput to the VM? If so, do I create an additional vswitch for one of the new Gigabit Ethernet ports?

     

    Any help will be greatly appreciated.

     

    Thanks.

    Blank General tab in Info Center

    in General Questions

    Posted

    In the process of trying out Jun's 1.03b bootloader on a new VM on my ESXI Server, I accidentally reused the serial number from another VM that was using the 1.02b bootloader. From other threads, I figured out how to edit the grub.cfg in the vmdk version of the bootloader to provide a new serial number and the VM booted but the Control Panel -> Info Center -> General page has no information on it

     

    Any suggestions for how to fix this will be greatly appreciated. (It is possible that this tab was blank when the original VM was created since I did not bother to look).

     

    P.S.: Synology Assistant does show the correct new serial number

    Capture.PNG

×
×
  • Create New...