Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register

Quick Links


Categories

In this Discussion

› › Need SolusVM Help
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Need SolusVM Help

lele0108lele0108 Member
edited March 2012 in General

Today, after around 3 weeks of waiting, I finally get my dedicated server (Intel 2x L5420, 8Gb EEC RAM). I install SolusVM on the server, and everything goes well, until I create a VM.

The VM refused to work, no matter what template, how I restarted it. I rebooted the server, and created a new VM, and it started to work. Yay.

I then proceeded to download OpenVZ templetes of the OpenVZ site. When I try to run these templates, the server REFUSES to start.

Can anybody help me? (I'm even willing to give admin details if somebody is that awesome!)

Comments

  • lele0108lele0108 Member

    I manually started this via SSH:

    dding IP address(es): 205.134.xxx.xx 205.134.xxx.xx
    Setting CPU limit: 400
    Setting CPU units: 1000
    Setting CPUs: 4
    Unable to start init, probably incorrect template
    Container start failed
    Stopping container ...
    Container was stopped
    Can't umount /vz/root/101: Device or resource busy

    That's weird. I got these directly of the OpenVZ wiki.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    are you sure you're on the right kernel?

    Francisco

  • lele0108lele0108 Member

    I am running SolusVM on CentOS release 6.2, 64-bit.

    I set all of the x64 templates as x64, and not as i386.

  • KuJoeKuJoe Member, Host Rep

    I think he means did you boot into the OpenVZ kernel (uname -a).

  • lele0108lele0108 Member

    Linux 191.cheetahhost.net 2.6.32-042stab049.6 #1 SMP Mon Feb 6 19:16:12 MSK 2012 i686 i686 i386 GNU/Linux

    Ahhhh. Er, how do we switch kernels?

  • prometeusprometeus Member, Host Rep

    You are using a 32 bit kernel....

  • lele0108lele0108 Member

    Gah, did my host REALLY install a 32-bit kernel when I told them to install a 64-bit one.

  • BassHostBassHost Member

    @lele0108 you want to post your grub.conf here?

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    Word of warning, if you're planning to sell any plans on these nodes, watch out for the .32 kernels :(

    Seriously consider RHEL 5 if you aren't using it for personal dev.

    Francisco

    Thanked by 1Raymii
  • prometeusprometeus Member, Host Rep

    @francisco what bad experience you had with.32 kernels?

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @prometeus said: @francisco what bad experience you had with.32 kernels?

    For a lack of a nicer way of saying it, it's unstable as fuck.

    Most nodes we put on it will go a couple days before it dead locks.

    Check the OpenVZ bugzilla to see all of the reports of softlocks/deadlocks/panic's on .32's.

    We tried to roll .32 and it was just a bad idea. We do .18's for now.

    Francisco

    Thanked by 1Raymii
  • prometeusprometeus Member, Host Rep

    Mmmmm I only have 3 nodes with openvz so maybe they aren't statistically relevant, but so far they are very stable (some weeks running) . Did you see the deadlock after some load pattern or it was random?

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @prometeus said: Mmmmm I only have 3 nodes with openvz so maybe they aren't statistically relevant, but so far they are very stable (some weeks running) . Did you see the deadlock after some load pattern or it was random?

    100% random. We got a few nodes (like 6) still on it that are stable, but most took a dive within a month. If we were OK with 30 day uptimes it'd be fine, but we're, including the clients, used to .18's 100 - 200+ day uptimes. :)

    Francisco

  • lele0108lele0108 Member

    @Francisco said: For a lack of a nicer way of saying it, it's unstable as fuck.

    Most nodes we put on it will go a couple days before it dead locks.

    Check the OpenVZ bugzilla to see all of the reports of softlocks/deadlocks/panic's on .32's.

    We tried to roll .32 and it was just a bad idea. We do .18's for now.

    Francisco

    Thanks for the heads up, I already talked to my host about switching my back to 64bit, didn't know why they provisioned me with 32. I am not selling this server, its for personal use :P

    @BassHost said: you want to post your grub.conf here?

    Thanks for the help, but I think I figured it out.

  • prometeusprometeus Member, Host Rep

    @Francisco said: 100% random. We got a few nodes (like 6) still on it that are stable, but most took a dive within a month. If we were OK with 30 day uptimes it'd be fine, but we're, including the clients, used to .18's 100 - 200+ day uptimes. :)

    ok, so I should cross fingers:

    10:13:20 up 25 days, 17:41

  • AsuraHostingAsuraHosting Member

    @prometeus said: so I should cross fingers

    That's what I'm wondering, haha; we just released OVZ's on a .32 as well.

  • miTgiBmiTgiB Member

    @prometeus said: ok, so I should cross fingers:

    Depends how you use it

    
[root@e3clt03 ~]# uptime
 08:17:22 up 56 days,  6:30,  
[root@e3clt03 ~]# uname -a
Linux e3clt03.hostigation.com 2.6.32-042stab044.11 #1 SMP Wed Dec 14 16:02:00 MSK 2011 x86_64 x86_64 x86_64 GNU/Linux

    It's a backup node, so only vsftpd, rsync and dropbear are running in 99% of the containers. No vSwap. No VPN's. With typical use is where @Francisco is running into troubles, so YMMV

  • prometeusprometeus Member, Host Rep

    @miTgiB said: vSwap. VPN's

    you named them all :)
    the nodes are in full production, one of the three (the first I installed 25 days ago) is near the limit of VPS I planned to stock on the servers for now...

  • prometeusprometeus Member, Host Rep

    @Francisco said: Most nodes we put on it will go a couple days before it dead locks.
    Check the OpenVZ bugzilla to see all of the reports of softlocks/deadlocks/panic's on .32's.
    We tried to roll .32 and it was just a bad idea. We do .18's for now.

    >

    I think to have found a pattern at least with softlock. Today a guy in the forum reported a dd test with very low performances. I checked and noticed I forgot to lower vm swappiness, which usually I set at a value of 1. While I was at it I checked the other servers and since on one I've a test vps on this server I set swappiness=0 to see if I could see some difference, after one minute load spiked out and I was flooded with softlock messages. Just setting it again to 1 stopped the mess and everithing went back to normal...

Sign In or Register to comment.