SCIENTIFIC-LINUX-USERS Archives

May 2014

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
James Rogers <[log in to unmask]>
Reply To:
James Rogers <[log in to unmask]>
Date:
Fri, 2 May 2014 08:34:27 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (51 lines)
arp could help.

But your best option is to use heartbeat / pacemaker, and deploy hard
fencing such as a networked PDU, or a managed switch.

On Fri, May 2, 2014 at 8:27 AM, John Lauro <[log in to unmask]> wrote:
> Any reason not just using heartbeat on the two nodes?  (yum install heartbeat)
> Then setup multiple interfaces between the two servers and let heartbeat bring services up/down as needed, and heartbeat can control which node has the secondary ip address active, etc.
>
>
> ----- Original Message -----
>> From: "Nico Kadel-Garcia" <[log in to unmask]>
>> To: "CS_DBA" <[log in to unmask]>
>> Cc: "Scientific Linux Users" <[log in to unmask]>
>> Sent: Friday, May 2, 2014 8:02:05 AM
>> Subject: Re: Network question - disable access to an IP without running ifdown ?
>>
>> Don't use the main IP address for your traffic. Use a secondary,
>> virtual IP address on the same port or ports, say "ifcfg-eth0:1", and
>> tie it to any of the common VIP monitoring packages such as is used
>> by
>> the master-master-mysql monitoring program. Basically, they
>> communicate over the available, primary IP address. If it's down, the
>> secondary VIP is kept down. When the primary address comes back down,
>> they negotiate over the primary address about which host to enable
>> the
>> secondary VIP on.
>>
>> On Thu, May 1, 2014 at 12:31 PM, CS_DBA <[log in to unmask]>
>> wrote:
>> > Hi All;
>> >
>> > we have a failover tool (for a database) that pings te master and
>> > if it's
>> > not responding runs ifdown over ssh to bring an ip alias down, then
>> > runs
>> > ifup of the alias on the standby node.
>> >
>> > A possible issue would be if the master was not reachable, thus the
>> > ifdown
>> > fails and once the master becomes available again then we have 2
>> > nodes with
>> > the ip alias and we risk a split brain issue
>> >
>> > Is there a way to disable any traffic from getting to the old
>> > master via
>> > commands we can run outside of the master server?
>> >
>> > Thanks in advance
>>

ATOM RSS1 RSS2