Howdy, Stranger!

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


How/Where do you host your website?
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.

How/Where do you host your website?

SimpleNodeSimpleNode Member
edited October 2012 in General

Currently all we have is a LoveVPS VPS running our main website, with hourly backups to our own nodes, a BlueVM VPS and a Hostigation VPS.

The problem right now is that if LoveVPS goes down, all we have are 3 backups, without a actual webserver.

How do you host your website, and what do you do when your webserver goes down? Do you have 2 webservers in sync, and DNS that failovers? If so, how do you sync them, and what DNS service do you use?

«1

Comments

  • jhjh Member

    Softlayer's Cloudlayer on SAN storage - so there's the failover.

    That's backed up every 15 minutes to 2x RAID6 NASs via R1soft.

  • I've got 2 webservers, one at Kazila and one at Prometeus. One is the master and the other is the slave.

    They get synced every 10 minutes using mysqlhotcopy and rsync.

    I've got a smaller VPS acting as a monitor and master DNS server (currently at BudgetVM). If it detects that the master is down, it starts a failover process and switches the DNS (TTL 300).

    The failover process also switches postfix, so email services failover, too.

    I run NSD for DNS on the 3 servers.

    I also run daily backups to a remote server using duplicity (monthly full backups, daily incrementals).

  • PacketVMPacketVM Member, Host Rep

    I host my main server on an Avante Hosting VPS, and it backs up to a Dewlance VPS every 5 minutes (although doesn's usually do that due to downtime). If that fails, I resort to my HostBluff VPS.

  • SimpleNodeSimpleNode Member
    edited October 2012

    But what if HostBluff is Up?

    And what if AutoBoot occurs?

  • PacketVMPacketVM Member, Host Rep

    @SimpleNode said: But what if HostBluff is Up?

    Well..that's a different story I guess. The Dewlance VPS has more downtime than uptime due to its 5-10 year old HDD's that the node runs and the old, rubbish hardware.

  • @dominicl Dewlance can't have downtime! AutoBoot will save the day!

  • InfinityInfinity Member, Host Rep

    I just moved my sites to a RAMnode (Atlanta), and host my WHMCS and some other stuff at a Hostigation (Charlotte) although I meant it to be the other way round but meh.. buy me some time. Backups are daily atm but that's prolly going to change and I'll probably put my non-DB driven sites on a hourly mirror and put some DNS magic to play.

    Thanked by 1Nick_A
  • letboxletbox Member, Patron Provider

    Our website hosted on our nodes colo in Dallas that is :).

  • PacketVMPacketVM Member, Host Rep

    @SimpleNode said: @dominicl Dewlance can't have downtime! AutoBoot will save the day!

    Haha, yeah true. Well, for real this time, I have a LoveVPS KVM in the UK and a LoveVPS KVM in Orlando. I've not completed the set up yet but the UK one will be used for the website and probably the Orlando one for backups. Not finalized yet ;)

    @Infinity has 2 brilliant hosts there!

  • PatrickPatrick Member
    edited October 2012

    Everyone so technical(basic setup)

    We run our VPS with corporate sites / blogs & other sites we own at KnownHost with cPanel running at 220MB RAM :) Currently using the dns / nameservers as default but we do have DNSMadeEasy which we will transfer soon for DNS Failover

  • PacketVMPacketVM Member, Host Rep

    @StormVZ said: KnownHost

    Are great :) It's a basic set up but from a class host!

  • AlexBarakovAlexBarakov Patron Provider, Veteran

    2 servers running in different locations... The more interesting thing is the cluster infront of them that filters the attacks at our main site :)

  • 24khost24khost Member
    edited October 2012

    Right now our main site is hosted on one of our vps in our South Bend location(Colostore DC).
    Billing is hosted with TurnkeyInternet on a XenVPS That I may have to reinstall to get some better lowend performance.

    DNS will be moving to a Seperate DNS server in Vegas and one In New Jersey and one in South Bend

  • KuJoeKuJoe Member, Host Rep
    edited October 2012

    We have 3 varnish load balancers set up in round robin DNS. Two are hosted internally on our network (one DDOS protected, one unprotected) and one is hosted with RamNode.

    All 3 load balancers cycle through 5 different webservers (1 protected internal, 1 unprotected internal, 2 with BuyVM, and 1 with Hostigation).

    In the event all 5 webservers are down (very unlikely) then the load balancers will continue serving cached pages. In addition to the redundancy, the performance has been amazing.

  • KuJoeKuJoe Member, Host Rep

    All for only $7/month. :)

  • I have a BuyVM, Future Hosting and Servint VPS, serving different websites.

    Some site are with Hawk Host and MDDhosting.

    DNS is handled by DME.

  • TazTaz Member

    On my pentium 3 personal PC with dial up connection

    Thanked by 3PacketVM netomx klikli
  • netomxnetomx Moderator, Veteran

    1 Chicago, and the backup in Neosurge. I have to test if my configs are correct:

    if chigado goes down, the other server will work. When Chicago goes up, it will get all the data from the slave and be the main server.

  • CloudFlare Business -> multiple vps/cloud instances (AWS 2 regions + our own onapp, that cross-check for HA)

  • Ash_HawkridgeAsh_Hawkridge Member
    edited October 2012

    Wow, some people have gone OTT with the fail-over haha. Our main site is hosted with Prometeus, using namecheaps DNS service and Google Apps email.

    WHMCS DB is backed up to a Gmail account, SolusVM DB is backed up to another off site VPS.

    Basically if something went wrong email would be up and we could move the site elsewhere in a few minutes if it was needed.

  • KuJoeKuJoe Member, Host Rep

    @GetKVM_Ash said: Wow, some people have gone OTT with the fail-over haha.

    You've never met Murphy have you? :P

    Thanked by 1netomx
  • @KuJoe said: You've never met Murphy have you? :P

    Luckily not yet, had my fair share of DDoS though :P

  • My personal website is hosted on an EDIS VPS up in Iceland, PointHQ DNS, backs up to Hostigation VPS daily.

    Hype.im is hosted on a Virt.IO VPS with Cloudflare in front + MaxCDN, files/MySQL backed up every 15 minutes to a Hostigation, which are then backed up to DropBox weekly.
    I'm currently working on making a WordPress Multisite installation have no SPOF, whilst being in sync at all times... without using Amazon S3 for the files :(

  • KuJoeKuJoe Member, Host Rep

    @GetKVM_Ash said: Luckily not yet, had my fair share of DDoS though :P

    Thankfully we've only had to deal with Murphy on occasion. I got tired of worrying about "what-ifs" so for $7/month I think it was worth going with an HA setup.

    Thanked by 2netomx Ash_Hawkridge
  • MaouniqueMaounique Host Rep, Veteran
    edited October 2012

    @KuJoe said: we've only had to deal with Murphy on occasion

    If Murphy really strikes, there is nothing in the world that can protect you. He will find a SPOF anywhere and even if there is none possible, you can still have a very determined kid with a ton of botnets against you taking down all the pillars of HA. I found out that not taunting Murphy and leaving room for failure in a controlled way is best, at least I know what failed when it happens.

    M

  • gubbytegubbyte Member
    edited October 2012

    I use an OpenVZ BuyVM. I haven't had any downtime yet, so I haven't needed to look into protection/other webservers. ClouDNS works fine.

  • I've had to deal with Murphy far too many times not to take reasonable steps

  • KuJoeKuJoe Member, Host Rep

    @Maounique said: If Murphy really strikes, there is nothing in the world that can protect you.

    Agreed. But I sleep better at night. There has been times where I'll wake up and see one of our webservers was serving Error 500 pages and was glad to know that for those 4-5 hours our clients and potential clients didn't notice. :)

  • Who is this Murphy?

Sign In or Register to comment.