Posts by cadre

    @aitrus00
    For what it's worth, i've run many Supermicro server boards, and one of mine behaves almost exactly like what you describe from a cold boot. This might be because this particular system doesn't use a Supermicro OEM power supply. However I also am using a Seasonic X750 (single 12V rail) on this system that's been running 24/7 for almost 8 years now.


    This system is loaded with eight 4tb 7.2k drives and eight 600g 10k drives with 32gb ECC and many add-in cards. So the cold boot start-up load is there.


    When I cold boot this system, I've found that after power on I'll have to press the RESET switch at least once (maybe twice) and then it'll continue to POST. Warm boots never have this problem.


    The only other thing I can suggest that you do over, is remove the CPU heatsink, re-apply minimal thermal paste, and make you that you DON'T over tighten the heatsink. It needs to be snug, and don't bottom out the threads. Try it, you may be surprised that it actually works...


    Good luck

    I can't find my reference now, but for what it's worth I know I have read somewhere about the Sandisk Cruzer Extreme USB drives using the same controller as there SSD's (with wear leveling) and seem to last much longer that most USB drives.


    Since then, I've been running a pogoplug with Debian Squeeze on one for well over a year now and no problems to speak of.


    You might consider trying one if you haven't already bought one, they are very convenient to use won't won't consume a SATA port.

    No, I don't believe that did make a change. The .161 system still doesn't show in the plugin. I'll see if I can "add" a hostname to either of those systems but I'm not sure if I can or not. the .161 IP system is a factory OS pogoplug and the .176 system is a APC SNMP network card.


    I appreciate the help and all that you've done, and I'll try to do some more testing. This is one of those services than I think many take for granted?


    Okay, sounds good. I don't have anything against the flashmemory plugin and will use it now that you've cleared that up for me. I would only use a USB install on a armel pogoplug system.


    Prior to stoneburner, the rrdcached service would consume 80 to 90% cpu utilization (constantly) and was pretty much unusable.


    But that's been fixed now. Thanks!

    I didn't realize that two or more system names with * would be the break down, but you're probably right. Although I'm not sure how I'll know ahead of time is any system on the network will fall into that category?


    I've installed the new package (thank you) and here's the output of "omv-mkconf dnsmasq"


    Question, I was under the impression that since Stoneburner has been released, that you can install OMV to a USB drive without worry if you disable "Monitoring" from the System section (without needing the flashmemory plugin)?


    As in disabling Monitoring reduces the disk writes, which also greatly reduces the system log entries as well.


    If the flashmemory plugin is still required for a USB system partition, then does "monitoring" still need to be disabled in Stoneburner too?


    Thanks,

    System info: This system is a 64-bit Wheezy VM running Stoneburner 2.1.4 with a 3.2.0 kernel.


    I've been using the dnsmasq plugin for probably close to a year now as my only DNS server on the network and have had no issues with DNS resolution (public or private).


    Recently I've started using the DHCP more and that's where I'm running into some issues. I've noticed that many times I'll have a newly powered on DHCP system that has an active lease, but that lease does not show in the plugin. I can list the contents of /var/lib/misc/dnsmasq.leases and find the IP of the system (if I know that MAC ahead of time), but it can get cumbersome with more systems on the network. I tried reducing the lease time to 1 hour in hopes of it trying to refresh and display all the leases, but that doesn't seem to make much difference.


    The first problem lead to the second, since all leases aren't always displayed I decided to make Static dhcp lease assignments. After every attempt to create a static lease, it would display an error message. Yet after logging out and back into the web-interface I would see the static lease that I tried to create.


    However when I tried to power on the system (of the static lease) it would be issued a different IP address. Usually one IP integer number greater than it was previously. I then tried creating the static lease from a current lease that was shown, but it was the same error and sequence as before.


    Lastly I tried to enable additional DHCP logging via the plugin, but that returned an error too.


    I don't believe I have anything misconfigured, or maybe I'm not entering the static lease information in correctly? I've been trying to figure this out, but right now just looking for any help or advice, but also curious if there's anyone else out there seeing the same thing?


    I've attached some pictures that hopefully will help.

    Update!


    I just found the problem on why I couldn't delete the drive from Shared Folders.


    Under Services, TFTP; it seems that this same data drive was listed as the active share path even though I've never activated TFTP.


    Once this TFTP shared folder was changed to "None" and saved, then I was able to remove the data successfully.


    I figured this was simple, but I looked and looked for the solution and just couldn't find it, especially since I never used TFTP.


    Thanks for the help.

    I don't believe that you have to worry about the raid volume itself, but just the alignment of the partitions themselves of course.


    Where possible I do prefer to create my raid volumes manually, but also realize that all debian should choose the correct values automatically?


    For any >2tb drive that I use in, I follow these short steps.


    at a terminal, using parted. (i.e. 3tb drive)


    Code
    parted /dev/sdX
    mklabel gpt
    mkpart primary 0 3001G
    align-check opt 1
    quit


    If the partition is aligned, you'll see that feedback. If not, you'll have to re-create that partition and adjust the size. for example in my experiences for creating a probably aligned partition for a 2tb drive with parted was simply "mkpart primary 1 2000G" So with that being said I wouldn't be surprised if your 4tb would be slightly different too.


    After all my partitions are created and aligned, then I'll proceed to create the Raid volume.


    Not to fill this post up, but if you want to read more about properly (or checking) aligned partitions please review these sites for references.


    Most information is redundant, but I doubt it would be a waste of time to review. Enjoy!


    Using Parted to create aligned partitions


    References
    http://rainbow.chard.org/2013/…performance-using-parted/


    http://support.hp.com/gb-en/document/c03479326


    http://blog.kihltech.com/2014/…tions-on-4k-sector-disks/


    http://stackoverflow.com/quest…alid-arguments-for-parted


    http://unix.stackexchange.com/…tion-aligned-using-parted

    Currently there is only one user account listed under Access Right Management and for the home directory field is disabled with Location and "None" under Settings.


    I'm not sure what exactly if anything is actually using this data mount.


    I'd really like to be able to fix this without re-installing everything from scratch.


    Any ideas, or do I have manually edit the config.xml to forcefully remove all references to this drive?


    Thanks for your help.

    I'm running OMV 1.19 on an ARM system. It's been working fine, but currently I wanted to remove a data drive. Seems simple but after disabling every service (except SSH) and UN-installing every single plug-in I still can not "delete" the shared folder associated with this drive.


    There are only two references to this drive's UUID in the config.xml file listed as below. I tried to edit the config.xml file but only ended up creating errors because I don't fully understand the language and formatting and had to revert the changes.


    What do I need to do to safely remove this drive from the system? I'm frustrated at myself because I can't figure this out. Please help.



    Code
    root@debian:/# cat /etc/openmediavault/config.xml | grep 8965d046-07cc-4c9f-ac8b-029bb33e4638
            <fsname>8965d046-07cc-4c9f-ac8b-029bb33e4638</fsname>
            <dir>/media/8965d046-07cc-4c9f-ac8b-029bb33e4638</dir>
    root@debian:/#

    Glad some of that worked. If some of your client stations are running windows, then you can modify the DNS settings so that any of your hosts configured through the DNS plugin will resolve without having to type in the ".lan" In your Windows client, modify the network adapter's TCP/IP setttings, then click the "Advanced" button, then under the "DNS" tab simply enter your workgroup name of "lan" in the field labeled "DNS Suffix for this connection". See attached, example of 2008 R2 server.


    In my network, OMV system is 192.68.1.24, and my gateway is 192.168.1.254. There are no devices on my network hosting any DNS or DHCP services except for the OMV system via the DNS/DHCP plugin. Every device with an internal NIC references the OMV system for DNS. This configuration works excellent for me.


    In my case my workgroup name is "BW" which in entered both under the OMV System, Network, General tab; as well as under the DNS plugin, Settings, General tab. Also here the only DNS server that I list is the OMV system.


    BUT, for the OMV system, Network, DNS Server tab, I entered two public (OpenDNS) servers. Internal and external DNS resolution works great, even on the OMV system of course.

    For starters, are you using the DNSMasq plugin v1.0.2 (that is the latest build to my knowledge)?


    Secondly, I see that under DNS Servers you entered both the OMV system and your router 192.168.1.2, I would remove the router IP.


    Also, I see that your NginX-Site system is not a DHCP member, so on that system have you manaully configured it's network to have a workgroup or setting of "lan"?


    Lastly, did you enter jukebox and 192.168.1.109 under the Static Entries tab section?


    Maybe unrelated, but under the OMV System, Network, General and DNS Server tabs, did you match the "Domain Name" of lan and what DNS Servers did you enter here?

    Thanks for the prompt reply, but I don't seem to have that problem. My current OMV system can resolve both local and public DNS names just fine from what I can tell; both from a terminal and via OMV web-interface.


    I just enabled and installed the OpenVPN plugin and downloaded a deb file from the webmin website (as a test).


    I send in a system report if you like, login ID is ibclhhTt


    Thanks,

    Hi David, could you clarify that a little more? I currently don't use any "internet" based services on the two OMV systems with the dnsmasq plugin, but so far the OMV systems seem to resolve both local and public DNS lookups correctly.


    Is there any specific example or test that would show this last issue to be solved before considering this plugin for the masses?


    Thank you all for your hard work and efforts.

    Thank you for the link, currently I have this installed on a 32bit VM (production DHCP and DNS) and on a ARMv5 pogoplug (testing) environment.


    Is there anything specific that you would like tested, or just time used and experiences with feedback?