Howdy, Stranger!

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


Hostigation LAX BGP Issues anyone? - Page 2
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.

Hostigation LAX BGP Issues anyone?

2

Comments

  • @KuJoe said: Move over to NC, it's nice over here. ;)

    I won't have power to add servers until Jan.

  • @cosmicgate said: It's not only you. I've been having these downtime since early feb/march and have since abandoned the vps on LAX.

    Well now this is something I am seeing become much more prevalent. I think it is due to the way I built the LAN, and before I start having the same issues as @Francisco is in San Jose, I've started moving to a much more segmented network. A jumbo VLAN1 was not such a wise choice, and there are now 1000's of hosts in it, so I have begun creating separate VLANs and will limit each to a /23 of IP space. The convenience of a massive IP pool is now lost with this. I should have the new servers ready to go over the weekend, with enough gear for a KVM node and a couple of OpenVZ nodes, since you've abandoned your VPS, you'd make an excellent test subject.

    Thanked by 1cosmicgate
  • I've never found Quadranet to be a great facility. Seems mighty popular in the VPS community though.

    Throughput every time I test something on their network underwhelms.

  • @pubcrawler said: I've never found Quadranet to be a great facility

    It is a better facility than most for the price

    @pubcrawler said: Throughput every time I test something on their network underwhelms.

    How does full available wire underwhelm?

  • Quadranet is then one of those price magnets. That makes sense.

    I've never seen full wire speeds to anything in Quadranet's facility. Could be the end resellers. Probably dealt with 3-4 different entities at that facility.

    Their own speed test file is rather pokey though too... Have tested it from dozens of other providers and have maybe two locations that get decent speeds, but far from gig speeds.

  • It's once again down.. (edit: just got up for now-.-)

    What should I expect to this sort of downtime? :3

  • I have my IRC bouncer on his LA nodes, ive never had an issue.

  • @klikli said: What should I expect to this sort of downtime?

    It will require a new IP, but put in a ticket and I will move you to a new VLAN. What is it you use to monitor?

  • @miTgiB said: What is it you use to monitor?

    me curious too....

  • I'll bet my bottom dollar it's UptimeRobot!

  • These are the outages my ping tracker has detected so far since I setup the service a week ago (the time is gmt+8).

    10/21/2012 19:04:15     00:01:01
    10/21/2012 19:07:58     00:02:12
    10/22/2012 03:02:19     00:03:23
    10/23/2012 01:14:17     00:09:49
    10/23/2012 01:30:12     00:01:01
    10/23/2012 14:44:32     00:02:12
    10/24/2012 10:57:04     00:02:33
    10/24/2012 11:01:28     00:01:12
    10/26/2012 17:20:52     00:02:03
    10/26/2012 17:54:29     00:29:24
    10/26/2012 18:27:20     00:07:56
    10/27/2012 01:58:04     00:01:00
    
  • flyfly Member

    ping detector

    mfw

  • klikliklikli Member
    edited October 2012

    It has been up and stable after the move ;)
    And, yes, it's UptimeRobot.

  • Mine is still giving random drops

  • @mitgib mine is down once again, can you please look into this (sorry on mobile can't log a ticket)

  • @klikli said: @mitgib mine is down once again, can you please look into this (sorry on mobile can't log a ticket)

    Hmm, just wondering, do you have some data starting from around 13/10? I am trying to figure out if it's the same dropouts I am getting from our LA servers as it sounds quite similar.

  • @victor deleted monitor on uptimerobot but I'll post my email notification times as soon as I can;)

  • @klikli said: @victor deleted monitor on uptimerobot but I'll post my email notification times as soon as I can;)

    Ah, alright no worries then. :)

  • @klikli said: can you please look into this

    Well, when I ping your VPS from home, it missed the 1st 8 packets, then not another after that, as if your nic was asleep. I adjusted the broadcast storm control setting in the switches, maybe they were blocking the arp packets?

  • @Victor said: the same dropouts I am getting from our LA servers as it sounds quite similar.

    Are you in pod 13? Both of my cabinets are on that floor, so we would be on the same switch for the uplink I think, maybe I'll force kate to join in here.

  • @pubcrawler said: Quadranet is then one of those price magnets. That makes sense.

    Quadranet has been more expensive than Colocrossing for all the quotes I asked about.

  • miTgiBmiTgiB Member
    edited October 2012

    @miTgiB said: maybe they were blocking the arp packets?

    --- 69.85.x.6 ping statistics ---
    1160 packets transmitted, 1160 received, 0% packet loss, time 1159580ms
    rtt min/avg/max/mdev = 56.970/57.181/63.780/0.240 ms
    

    Since the adjustment, I am not getting the flood of solus errors and this is to your VPS @klikli I'm leaving a ping going from home and will check it every so often today, but I have my fingers crossed for now.

  • That's great, though I seriously doubt about that since uptimerobot it's pinging it every five minutes..

  • Chris_QNChris_QN Member
    edited October 2012

    In response to "pubcrawler"s statements:

    1. That pubcrawler guy sounds fishy how he jumps in here with negative comments.

    2. your statements about price are complete bull. we are not cheap, we actually value our service a lot higher than "others" here in the building and do not drop our pants just to make a sale. compare your provider's advertised prices to ours... and we are well worth it.

    3. our network is EXTREMELY fast. we have only the best providers that we've chosen through years and years of testing and experience; not to mention the continual route adjustments we make to lower latency and increase speeds. we continually get compliments on our network and you (surprise) are the only one complaining -- and surprise, you say its via "OTHERS" on our network, not direct service from QuadraNet.

    4. we have 24/7 on-site support that know what they are doing and take care of our customers.

    People choose us because we are good and their clients want to be on our network and in our datacenters. Simple as that, and we know it.
    We base our growth on QUALITY not price. We ensure our service is good so that people want to be here FOR THE SERVICE, not for the price.

    You should also do some research as to how many providers who have "LA locations" are actually IN one of our datacenter pods and/or using our network. Same goes for our Dallas location (and as Miami grows, I am sure Miami...)

    • Chris G @ QuadraNet
  • Shit just got serious

  • @klikli said: That's great, though I seriously doubt about that since uptimerobot it's pinging it every five minutes..

    I keep thinking it has something to do with the arp table the way this has been acting, so looked into that more today, and compared it with how Charlotte doesn't have this error, and this only started recently, a little google sevantism and I had my idea how to test this. I quadrupled the size limit of the arp table in Los Angeles, and after waiting a few, it looks like this was a good idea.

    tim@escapefromla:~$ cat /proc/net/arp | wc -l
    1494
    

    The default on a Vyatta is 1024 for arp table size limit, increased to 8192, so if you've been experiancing the random drops, try to keep a closer eye and report back.

  • If you have arp table issues you would be seeing a lot of "neighbor table overflow" messages on dmesg (no idea if you can do this on vyatta).

    Thanked by 1miTgiB
  • miTgiBmiTgiB Member
    edited October 2012

    @rds100 said: you would be seeing a lot of "neighbor table overflow"

    Cool, we got to the same point, taking different roads


    [1352837.175658] ipv4: Neighbour table overflow.
    [1352837.457266] ipv4: Neighbour table overflow.
    [1352837.493654] ipv4: Neighbour table overflow.
    [1352837.915238] ipv4: Neighbour table overflow.
  • Ok, so you have the right suspect :) If you get rid of these messages everything should be OK.

Sign In or Register to comment.