Re: LVS backup router not taking role

Ranjith Rajaram <[email protected]> Wed, 18 Jan 2012 19:29:55 +0530
Newsgroups gmane.linux.redhat.piranha
Message-ID <[email protected]>
This is a multi-part message in MIME format.
--===============2384427695232912635==
Content-Type: multipart/alternative;
	boundary="------------020002080905070503000203"

This is a multi-part message in MIME format.
--------------020002080905070503000203
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit



On 01/18/2012 07:22 PM, Khapare Joshi wrote:
> Hi, all,
>
> first time poster. I recently deployed lvs (primary and backup) with 
> piranh and ran into few issues. I turned off the primary lvs to see 
> backup lvs will take over. However it does not take over right a way 
> seem require restart pulse. How does this fail over works if primary 
> goes down ? should not this be happen right a way if primary lvs goes 
> down ?
If primary goes, secondary has to take it over

It depends on deadtime

  deadtime = n
               Length of time before a node is declared dead and IP 
takeover occurs.

lvs.cf needs to be checked together with system configuration.  Look at 
/var/log/messages




>
> And I have four real server- and it sees if one goes down whole system 
> is unavailable (httpd service). Then I restart pulse and get  the 
> httpd service. I do not think this is right. could be please guide me 
> in right direction to make reliable.
>

Do you have separate boxes for LVS directors and real server's ? [Are 
you running pulse on any of the two real server's ?]

check the output of ipvsadm -L -n from the active lvs server [paste it here]


> PS: I use direct routing and have
>
> net.ipv4.conf.lo.arp_ignore = 1
> net.ipv4.conf.lo.arp_announce = 2
>
> in /etc/sysctl.conf
>
> sorry for too many question.
>
> Thanks
>
>
> _______________________________________________
> Piranha-list mailing list
> [email protected]
> https://www.redhat.com/mailman/listinfo/piranha-list


--------------020002080905070503000203
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    <br>
    <br>
    On 01/18/2012 07:22 PM, Khapare Joshi wrote:
    <blockquote
cite="mid:CABHhH+sbBaQzJVXKm4Zg59UiSj5TvnV3eGvJOFB51e3XfD5fKQ@mail.gmail.com"
      type="cite">
      <meta http-equiv="Context-Type" content="text/html;
        charset=ISO-8859-1">
      Hi, all,<br>
      <br>
      first time poster. I recently deployed lvs (primary and backup)
      with piranh and ran into few issues. I turned off the primary lvs
      to see backup lvs will take over. However it does not take over
      right a way seem require restart pulse. How does this fail over
      works if primary goes down ? should not this be happen right a way
      if primary lvs goes down ?<br>
    </blockquote>
    If primary goes, secondary has to take it over<br>
    <br>
    It depends on deadtime<br>
    <br>
    &nbsp;deadtime = n<br>
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Length of time before a node is declared dead and IP
    takeover occurs.<br>
    <br>
    lvs.cf needs to be checked together with system configuration.&nbsp; Look
    at /var/log/messages<br>
    <br>
    <br>
    <br>
    <br>
    <blockquote
cite="mid:CABHhH+sbBaQzJVXKm4Zg59UiSj5TvnV3eGvJOFB51e3XfD5fKQ@mail.gmail.com"
      type="cite">
      <br>
      And I have four real server- and it sees if one goes down whole
      system is unavailable (httpd service). Then I restart pulse and
      get&nbsp; the httpd service. I do not think this is right. could be
      please guide me in right direction to make reliable.<br>
      <br>
    </blockquote>
    <br>
    Do you have separate boxes for LVS directors and real server's ?
    [Are you running pulse on any of the two real server's ?]<br>
    <br>
    check the output of ipvsadm -L -n from the active lvs server [paste
    it here]<br>
    <br>
    <br>
    <blockquote
cite="mid:CABHhH+sbBaQzJVXKm4Zg59UiSj5TvnV3eGvJOFB51e3XfD5fKQ@mail.gmail.com"
      type="cite">PS: I use direct routing and have <br>
      <br>
      net.ipv4.conf.lo.arp_ignore = 1<br>
      net.ipv4.conf.lo.arp_announce = 2<br>
      <br>
      in /etc/sysctl.conf<br>
      <br>
      sorry for too many question.<br>
      <br>
      Thanks
      <pre wrap="">
<fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
Piranha-list mailing list
<a class="moz-txt-link-abbreviated" href="mailto:[email protected]">[email protected]</a>
<a class="moz-txt-link-freetext" href="https://www.redhat.com/mailman/listinfo/piranha-list">https://www.redhat.com/mailman/listinfo/piranha-list</a></pre>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">
</pre>
  </body>
</html>

--------------020002080905070503000203--


--===============2384427695232912635==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Piranha-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/piranha-list
--===============2384427695232912635==--