nv-l
[Top] [All Lists]

Re: RES: [nv-l] Question#1 - Redundant & Disaster Recovery Solution f or

To: nv-l@lists.tivoli.com
Subject: Re: RES: [nv-l] Question#1 - Redundant & Disaster Recovery Solution f or (AIX) Net view Server
From: Stephen Hochstetler <shochste@us.ibm.com>
Date: Wed, 12 Mar 2003 18:01:08 -0600
Delivered-to: mailing list nv-l@lists.tivoli.com
Delivery-date: Thu, 13 Mar 2003 00:12:02 +0000
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
List-help: <mailto:nv-l-help@lists.tivoli.com>
List-post: <mailto:nv-l@lists.tivoli.com>
List-subscribe: <mailto:nv-l-subscribe@lists.tivoli.com>
List-unsubscribe: <mailto:nv-l-unsubscribe@lists.tivoli.com>
Mailing-list: contact nv-l-help@lists.tivoli.com; run by ezmlm



Yes,  Tivoli supports an HACMP solution.  The installation is documented in
the NetView documents.  This is a good hardware high-availbility solution.

For those that want to use 2 boxes, using the PAX/TAR of the database to
the second machine is a very acceptable practice.  NetView development has
seen enough people doing this that in V7 there are new ways to make the
database "quiet" to make a backup without taking all the daemons down.
This makes your backups less disruptive.

The question comes, how fast does failover have to occur?  How do you
handle trap processing?  TEC Integration?  How do I keep configuration
changes in sync with my backup NetView.  Those are issues that several of
us have tackled over the past several years with customers.

If your failover needs to happen faster than NetView can start on your
backup server, then the second NetView server has to be running.   So after
the database is restored to the backup, NetView actually needs to be up and
running.   I have architected solutions that automatically control the flow
of events from both NetViews to TEC.  If primary NV is up, the second does
not send events.  When the second sees the primary NV application down, the
second starts sending events.  This can be automated.

For trap processing, I recommend that all traps be forwarded to a MLM.  I
know, it can only handle SNMPV1 traps.   If you have SNMPV2 traps you may
just have to configure your network devices to send traps to both NetViews.
If your devices can only send to one address and you cannot use MLM, then
this is the case where an HACMP solution is required.   When both primary
and secondary NetView discover the MLM, they put themselves in as trap
destinations.   This allows devices to send to one address, the MLM, but
both NetView's receive the trap.  If are HAVE to have a redundant MLM to
remove a single point of failure, then you want to configure a second
machine with either IP address takeover or simply power off primary and
power on redundant MLM machine which has the same IP address and MAC.

HACMP uses shared disks between machines.  The configuration recommends
that you put the NetView database on the shared disk.  But if you have a
software glitch due to some rogue SNMP agent in your network the result
could be toasted NetView database.  HACMP would not help you.  If you had a
primary / backup server you would have a solution for that problem.   For
the HACMP you could extend that solution by taking a database backup
regularly and stored that on tape or another space, then you would have the
best of both worlds.  If your NetView dies due to database issues, restore
from last good database and get running.  If your NetView dies due to
hardware, fail over to the other HACMP node.

One note, two NetView's don't really add that much traffic to the network.
Putting two addresses in all your trap destinations does add a lot of
traffic.  It depends how you are configured, but it always seemed like
NetView was getting more traps than what people know what to do with.   A
trap is a whole lot more data than an ICMP packet.  That is another reason
that a MLM as trap receiver is attractive.

Steve Hochstetler


>"Ken Karasek" <ken.karasek@hewitt.com> writes:
>
>> We taken a less costly and less complicated solution after looking
>> into the HACMP options listed by Marcos below..  The question that
>> will need to first answer is the recovery time required to have your
>> redundant NV/6000 online and if the NV/6000 recovery has to be fully
>> automated.
>
>Curious--does Tivoli support an HACMP solution anymore?
>
>We use two separate NetViews that are up continuously--only one of
>which is forwarding events to TEC at any given time.

Kind regards,
Stephen Hochstetler              shochste@us.ibm.com
International Technical Support Organization at IBM
11400 Burnet Road   Austin, TX  78758
Office - 512-838-6198 (t/l 678)       FAX - 512-838-6931
------------------------------------------------------------
http://www.redbooks.ibm.com




---------------------------------------------------------------------
To unsubscribe, e-mail: nv-l-unsubscribe@lists.tivoli.com
For additional commands, e-mail: nv-l-help@lists.tivoli.com

*NOTE*
This is not an Offical Tivoli Support forum. If you need immediate
assistance from Tivoli please call the IBM Tivoli Software Group
help line at 1-800-TIVOLI8(848-6548)


<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web