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

› › Breezehost is going to start KVM Virtulization :)
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.

Breezehost is going to start KVM Virtulization :)

BreezehostBreezehost Member
edited April 2012 in General

Hello Lowendbox members or lovers

We are glad to start KVM Virtulization!!!! I just wanted to know or need some suggestions from you regarding server location. Can you please tell me that which location is the best in case of KVM like USA,Germany,France,UK,Europe or any other which you think better. Please suggest.

Thanks,

«1

Comments

  • BreezehostBreezehost Member

    Okay liam Thanks, so it's better to go for East Coast or UK or France or texas right ?

    Which data center is good ?

  • BreezehostBreezehost Member

    @DotVPS : so it okay to go for UK location ? and what you think about Germany ?

    Thanks,

  • subigosubigo Member

    This makes what.. ten "which location should we choose next" threads this month?

    Thanked by 2[Deleted User] Spirit
  • BreezehostBreezehost Member

    @liam and DotVPS: Can you suggest me many more 2-3 locations and data center ?

  • BoltersdriveerBoltersdriveer Member, LIR

    Would be nice if you could get into Asia-Pacific ;) . Also, OVH for France but... eh. Not really sure what else.

    Thanked by 1NanoG6
  • FranciscoFrancisco Top Host, Host Rep, Veteran
    edited April 2012

    @Breezehost said: @liam and DotVPS: Can you suggest me many more 2-3 locations and data center ?

    Isn't that your job as a host? To do research and find a market you like?

    Rules of the LEB market:

    1) Know a market that you want to target. Buying boxes everywhere leads you to #2
    2) Stay out of the deadpool
    3) STAY OUT OF THE DEADPOOL

    You've broken all 3 so far.

    Francisco

    Thanked by 1Liam
  • tuxtux Member

    @DotVPS said: Not Ecatel like you currently use for vz ;)

    Ecatel is OK.

  • BreezehostBreezehost Member

    @Boltersdriveer @DotVPS and @liam : Thanks for your suggestions. Now, lets have a discussion about configuration.. what you think how much Bandwidth or RAM size, RAID etc.

    @Francisco : we have already one discussion on one post before some days that we are not deadpool. Please don't say like this. Thanks,

    Thanks,

  • Ash_HawkridgeAsh_Hawkridge Member

    Stay away from UK KVM, that's our patch (And maybe loveVPS/EDIS) LOL

  • BreezehostBreezehost Member

    @DotVPS : Yes I want you to suggest location but need some suggestions on configuration also. As you already know discussion is always good if you are going for latest technology. We know the configuration but it's always good to take guidance

  • JacobJacob Member

    I don't think the 4TB is really neccessary, 1 / 2TB on VPS Nodes is usually enough.
    4 x 1TB / 2 x 1TB.

    @DotVPS said: HW RAID10 4 x 2TB SATA disks

  • Ash_HawkridgeAsh_Hawkridge Member
    edited April 2012

    Actually DotVPS's config is bang on for KVM. Apart from the gbit port minimum :P

    The hardware RAID card is the most important, do not use software RAID with KVM, trust me.. And stay away from CentOS 6.

  • TaylorTaylor Member

    @VMPort said: gbit port minimum

    I was thinking that to, it would cost crazy money to go over a gbit if your doing it a node at a time.

  • BreezehostBreezehost Member

    We will offer 1TB/1.5TB/2TB for a VPS on our node so on an average if we are going to host max. 40 KVM's then it will definitely require upto 50TB bandwidth. right ?

    @VMport : why you are advising to stay away from CentOS 6 because we are comfortable with CentOS 6.2 + KVM and you already know the CentOS is best..

  • Ash_HawkridgeAsh_Hawkridge Member

    @Breezehost said: @VMport : why you are advising to stay away from CentOS 6 because we are comfortable with CentOS 6.2 + KVM and you already know the CentOS is best..

    You are already comfortable? You don't even offer KVM yet do you?

    CentOS is the best, CentOS 6 is not when KVM is concerned. I'm sure @Francisco can jump in and throw some knowledge your way. I'm tempted to let you find out for yourself :P

  • prometeusprometeus Member, Host Rep

    @VMPort said: And stay away from CentOS 6.

    Did you have bad experience with this? As far I can see our RHEV 3 cluster (based on rhel 6) is stable as it was vers. 2 (rhel 5)...

  • Ash_HawkridgeAsh_Hawkridge Member
    edited April 2012

    @prometeus said: Did you have bad experience with this? As far I can see our RHEV 3 cluster (based on rhel 6) is stable as it was vers. 2 (rhel 5)...

    We ran CentOS 6.2 on a box the other day and could not pull any more than 30MB/s with throughput/DD tests. We tried virtio and IDE drivers, different boot params etc but nothing could get the tests above that rate within a KVM-Domain.

    We initially thought that it may be software RAID10 (Which probably did have a slight effect) but the node itself was pulling 240MB/s. When we looked further into it we found a lot of issues regarding this with RHEL6.

  • BreezehostBreezehost Member

    @VMPort : See this is not the problem of CentOS this is depend on server performance, server configuration. You just need to check VirtIO as well as BenchmarkIO performance then you will not face the problem.

    We have already tried this practically with success and now going to do live.

  • Ash_HawkridgeAsh_Hawkridge Member

    @Breezehost said: @VMPort : See this is not the problem of CentOS this is depend on server performance, server configuration. You just need to check VirtIO as well as BenchmarkIO performance then you will not face the problem.

    We have already tried this practically with success and now going to do live.

    Ok i must be wrong then.

  • BreezehostBreezehost Member

    @VMPort : just want to clear your views and the discussion is only for what is good and what is bad so but thanks for your help also :)

  • BreezehostBreezehost Member

    All my dear friends,

    Still waiting for good suggestion from you about location as well as configuration.

    Thanks all,

  • BoltersdriveerBoltersdriveer Member, LIR

    Pssh. Configuration? As a host, you really need to know what you need, what your clients need, and etc. I'd HIGHLY recommend you get a HW RAID card. Not getting one easily leads to #2 that @Francisco pointed out above.

  • prometeusprometeus Member, Host Rep
    edited April 2012

    @VMPort maybe it was some nasty offset effect. Did you try only one configuration/server?

    Edit: sorry for the OT

  • Ash_HawkridgeAsh_Hawkridge Member

    @prometeus said: @VMPort maybe it was some nasty offset effect. Did you try only one configuration/server?

    Edit: sorry for the OT

    Maybe, this was isolated to one server using the same settings and params we use across our other servers. The only difference been software RAID10 and RHEL6 instead of 5.7.

  • LV_MattLV_Matt Member

    @VMPort said: Stay away from UK KVM, that's our patch (And maybe loveVPS/EDIS) LOL

    My patch actually :)

    @VMPort said: The hardware RAID card is the most important, do not use software RAID with KVM, trust me.. And stay away from CentOS 6.

    Spot on, you can't afford to cheap out on KVM.

  • Ash_HawkridgeAsh_Hawkridge Member

    @LV_Matt said: My patch actually :)

    I included you, you should be grateful haha :P Bit of joint advertising right there.

  • miTgiBmiTgiB Member

    @VMPort said: We ran CentOS 6.2 on a box the other day and could not pull any more than 30MB/s with throughput/DD tests. We tried virtio and IDE drivers, different boot params etc but nothing could get the tests above that rate within a KVM-Domain.

    [root@e3clt01 ~]# cat /etc/redhat-release
CentOS release 6.2 (Final)
[root@e3clt01 ~]# w
 10:12:49 up 80 days,  2:20,  1 user,  load average: 2.82, 2.77, 2.48

    Then from within a VM on the node, while not blazing, certainly better than 30M/s

    [root@dns3 ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 13.4745 s, 79.7 MB/s
[root@dns3 ~]# cat /etc/redhat-release
CentOS release 6.2 (Final)

    I can reproduce this and have seen many others produce numbers like this or higher on other EL6 KVM nodes, so you must of hit the wrong button on something.

  • Ash_HawkridgeAsh_Hawkridge Member

    @miTgiB said: I can reproduce this and have seen many others produce numbers like this or higher on other EL6 KVM nodes, so you must of hit the wrong button on something.

    Yep, we have come to the conclusion that something was conflicting with CentOS 6 and our hardware set up. We will be happy to stick with 5.7/5.8 until it gets sorted.

    I still would heavily advise against not running software RAID with KVM though :P

  • birdie25birdie25 Member

    @Breezehost try qirus.eu, they work out of Serverius :)

    Thanked by 1TheHackBox
  • AldryicAldryic Member

    @Breezehost said: we have already one discussion on one post before some days that we are not deadpool. Please don't say like this. Thanks,

    @DotVPS said: He has a valid point.

    Not really. In the same thread he's referencing, I posted several links to threads announcing/covering Breezehost's prior deadpool status.

    @Breezehost said: Thanks for your suggestions. Now, lets have a discussion about configuration.. what you think how much Bandwidth or RAM size, RAID etc.

    Seriously? Are you just hurting for a business model that's sustainable, or do you honestly have no clue what you're doing?

    Thanked by 1DeletedUser
Sign In or Register to comment.