Server redundancy

Hi!
I'm currently working on a router based on Arch.
The thing that I want to accomplish is to have two machines working redundantly.
They've got the same hardware etc, so the deal is to keep the configuration files up2date (updated through various scripts dynamically), and when one machine dies, the other should take over, having updated configuration files and, obviously, the same IP as the first one.
Anyone got hints where to start out? Had a look at keepalived.org, seems promising (except for the 1+ year silence), but not sure how well it works.
I'm looking for a solution that will also work for other types of servers, databases, www, dns, mail etc etc... if possible.
Any ideas, tips, hints are welcome : )
/Diddi
Last edited by diddi (2009-01-20 20:18:46)

lets see...
relayd+carp (openbsd)
proprietary vrrp (cisco stuff)
heartbeatd
keepalived (lvs, linux-ha)
using a frontend proxy like ha-proxy (layer 3/4 or layer 7)
whackamole (build on spread toolkit)
or a proprietary frontend LB like coyotepoint or fat-pipe (h5 i think)
a portable ip block with multiple routes broadcast through BGP (think same ip in different locations). often used with dns servers (opendns does this for instance)
likely you want a combination of some components of the above....
Last edited by cactus (2009-01-23 18:02:26)

Similar Messages

Bpel server redundancy

Hi!
in our product environment now we have 1 bpel pm server 10.1.2 and we'd like to have one more redundant server for this instance, which should work in case of crash of the first server ...
Could you point me to the right documentation about this problem? thanks a lot,
Tomas

Looks like the server is unable to connect to the dehydration database.
A standard installation would have this data source setup as part of the installation.
"jdbc/BPELServerDataSourceWorkflow"
This again uses the connection pool "BPELPM_CONNECTION_POOL"
And the name of the data source is "BPELServerDataSourceWorkflow"
There are a few checks that you can perform.
1. Check if the dehydration database is up.
If no then this should be up and running.
2. Go to the EM and use {test connection} for "BPELPM_CONNECTION_POOL"
If comes back with error then check if the dehydration database credentials has changed post installation.
There are a few other datasources that share the same connection pool. Check if these are working too.
Every Little Helps
Kalidass Mookkaiah
http://oraclebpelindepth.blogspot.com/

Poor man's server redundancy?

I am helping a couple of friends who have small businesses, each with a single server. They have become concerned about the business cost of a server outage, but have a limited budget for redundancy. I used to work on this stuff, but have been
out of it for several years, so I'd like to solicit suggestions from you all. Our goals are to provide quick & simple server failover at a modest cost, but not necessarily instantaneous or automatic. I've thought of a few categories of solutions:
1. Off-line server spare - Keep another server of same HW configuration; if live server, which has pair of HDD's in RAID-1 mirror, fails, remove one of the drives from the failed server, pop it into the spare server and fire it up, with same server name
& IP address. I wonder if this would work - the disk controller in the spare server might not accept a drive from the other server. Also this would not protect from a major electrical problem, say both drives got zapped.
1a. A variation of server spare - maintain a spare server with distinct name & IP address, synchronize data nightly. If live server fails, shut it down, then rename the spare server and give it the IP address of the failed live server.
2. Failover SW such as Double Take or StorageCraft.
3. Server cluster - probably too expensive, especially if a SAN were used.
4. Virtualization.
Appreciate any suggestions.
Bob

For simplicity, have them use a virtual server. This can be moved easily to a new server because the virtual hardware is identical. Moving a physical disk to a spare server is going to be more problematic, because there may be hardware differences. Also,
the network hardware will have different identifiers, so network configuration will have to be re-done etc. This is not insurmountable, but still could be more difficult than it has to be.
Moving a virtual machine is as simple as copying it and starting it again. Of course, you still need to backup the data regularly to prevent against accidental deletion.

Directory Sync server redundancy.

How is redundancy for the Directory Sync server accomplished?
We have a load balanced pair of ADFS servers, a load balanced pair of ADFS proxies at our primary datacenter. We have additionally deployed an individual ADFS server and proxy at an alternate datacenter. The ADFS servers are configured as a farm. So if we
need to fail over to the secondary datacenter, we just need to change DNS entries for our federation services. All this is in place and tested.
How do we accomplish something silimar for Directory Sync?

Implementing FIM is another investment in terms of licensing, I would rather go with DirSync ;)
I have heard that Microsoft is planning to get rid of DirSync in future. I think I will be good with Standalone servers for a while.
Cheers,
Gulab Prasad
Technology Consultant
Blog:
http://www.exchangeranger.com    Twitter:
  LinkedIn:
   Check out CodeTwo’s tools for Exchange admins
Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

Is it possible to make the ISE guest server redundant ?

Hi,
We've an ISE cluster of two ISE nodes.
The ISE guest server works fine on the primairy ISE node.
MAC address of the guest client is set in the map 'GuestDevices' after accepting the AUP policy.
The the ISE sents the COA and the client authenticates again and is punt in the guest vlan.
But when the primairy ISE is offline, I see the guest portal AUP page on the secondairy ISE node.
I can accept the AUP policy, and I get an error message.
On the secondairy ISE I see that the COA to the switch is sent, to clear the session to the primairy ISE....
But the COA request should ask to clear the session to the secondairy ISE ( the primairy ISE is offline ).
Should it be possible to configure the ISE guest functionality redundant in an ISE cluster?
/SB

The Guest portal can run on a node that assumes the Policy Services persona when the primary node with Administration persona is offline. However, it has the following restrictions:
•Self registration is not allowed
•Device Registration is not allowed
•The AUP is shown at every login even if first login is selected
•Change Password is not allowed and accounts are given access with the old password.
•Maximum Failed Login is not be enforced
http://www.cisco.com/en/US/docs/security/ise/1.0/user_guide/ise10_guest_pol.html#wp1126706

DAGs and Edge server redundancy

We are in the process of creating an Exchange DAG and setting up redundancy between our edge servers and need to know if there will be any type of service outage as a result of any of these changes? If so, how should we expect the process to go?
E.G., will there be a service outage after the DAG is created but not after the edge redundancy is set? etc.

Hi sbusarow,
For edge server:
You can deploy multiple Edge Transport servers and use multiple DNS MX resource records to load balance activity across those servers. You can also use Network Load Balancing (NLB) to provide load balancing and high availability for Edge Transport servers.
For DAG:
You could refer to below:
http://technet.microsoft.com/en-us/library/dd638121.aspx
So, if you understand how to achieve the target, you will get the information.
Regards!
Gavin
TechNet Community Support

Quorum Server Redundancy Question

Hi All,
I'm just investigating my options for a new cluster configuration and was trying to find out about multiple quorum servers hosts. All the examples I have come across in the documentation have 1 physical host acting as a quorum server for an n+1 node cluster. I'm ssuming that there will be quorum issues in the event the physical host, hosting the quorum server, is down and the cluster nodes performed a reconfiguration/switch when the quorum server was unavailable.
Is it possible to have 2 physical hosts, with quorum servers defined on each, that can then be confidured in to the cluster. Effectivly pointing at two different quorum servers for votes?

Correct, the QS is only used if the cluster changes state, i.e. nodes leave or join. However, having more than 1 QS for a single cluster does not help. You simply lower your overall availability because there are more failure scenarios where one of these is down, leading to insufficient votes for the remaining cluster node to obtain.
Active monitoring and prompt repair of the QS (or QD) is the right approach.
Tim
---

ICloud Server Redundancy

There's very little information out there about how iCloud works behind the scenes. I hear about the giant datacenter in North Carolina, as well as others. I certainly expect data storage on RAID and networks and servers arranged so no single device or hardware failure should cause a problem. But what if something more catastrophic happens? A single natural disaster in North Carolina could render all kinds of local redundancy irrelevant. Does anyone know if we have any actual geographic diversity on our iCloud data?

I never said anything was anyone's fault. Try reading. Of course the servers should not be down for 2 days. Of course it should work properly. You think Apple planned it like that? But clearly there was a major problem that took considerably longer to fix than anyone, including Apple, expected. None of us here know what exactly went wrong. Perhaps most of us wouldn't even understand it if Apple described the problem in great detail. Fixing networks and servers is not as simple as fixing the tyre on a car.
These things happen. GoDaddy (one of the worlds largest webhosts) had an outage some days ago that took down milions of commercial and personal websites and email. No technology company is immune to such technical failures. Sometimes they take considerable time to identify the problem and further time to fix it.
If you are not using a free, consumer-level email service for business purposes, then my comment does not apply to you.
However, if you need an email service that gives guaranteed service levels, and guaranteed uptime, use one. There are plenty out there. iCloud is not that service.

Redundant Portal Server

Hi, we are doing some tests and we would like to have 2 machines running the portal and one gateway at the front. The first thing we are trying is to make Identity Server redundant.
So we installed on Machine A the identity Server and then in Machine B, added machine B to the server list in the platform service configuration. If I log in to machine A I can switch to machine B without problems, but if I try to log in directly into machine B I can't and I get the following message:
Authentication Service is not initialized.
Contact your system administrator.
Any ideas or suggestions on how to solve it and how to have a redundant identiy and Portal servers?
Thanks

I've got the same problem and how I revolved it.
I copied AMconfig.properties and serverconfig.xml from the working server to the not working one. I change the configuration file for changing the server name and restarted the service. Everything works fine now. The thing is the password inside serverconfig.xml is encrypted with a key in AMConfig.properties. If you want your server to work, they need to have the same key and password.
Hope this help...

Quick question: redundant OMS databases for Grid Control

Dear Friends,
I am wondering if it is possible to configure redundant OEM Management Server databases on separate servers and how would that be done?
I don't think it is possible to have an agent write to more than one OMS database, so how is it possible to have our OEM management server redundant?
Is running DG the only option?
Thanks.
Ji Li

Damorgan:
I thought I remembered doing this at a previous site that I worked a few years ago, but I don't remember how I did it.
Will I need to configure dual listeners, or use separate ports for the dual agents?
Do you know where there might be instructions for how to set up and run dual agents on the same server?
Thanks again.
Ji Li

Paging server

Hello I purchased 2 business edition 6k one as sub and the other as pub , each one has basic paging server license , how can i implement paging server redundancy without need to purchase the advanced notification ?

Hi,
As per the following the Advanced license key is required for Informacast resiliency
https://www.singlewire.com/help/advanced/cucm/Administer_the_Virtual_Appliance/Resiliency/Configure_CUCM_for_IC_Resil_and_Install_VA.htm
HTH
Manish

Design for DHCP redundancy.

Hi,
Can somebody advice , how can i have dhcp server redundancy in my network. I need to design a network that will have around 30 vlans spanned across 50 buildings.

Hi Renzil,
I'm actually working on a similar solution at work at the moment. There are a couple of options (some have mentioned already)
1. Have two dhcp servers in different locations, with the same scopes and settings configured. Add the IP helper address for both servers to any device that needs to relay dhcp broadcasts. Then choose either;
a) Active/Standby - Configure the dhcp servers to act as a Primary/Secondary, by disabling all scopes on the secondary server until a situation arises where you need to failover. When you failover, enable the scopes on the secondary dhcp server. Be careful to prevent the Primary from giving lease when it comes back online.
b) Active/Active - Split the scopes across both dhcp servers in a 20/80 or 30/70 configuration as mentioned above. You will need to ensure the same reservations are configured on each dhcp server.
c) Active/Active - Disable some scopes on one of the servers, and have the same scopes active on the other server. When failover occurs, enable the scopes that are disabled on the server that is still reachable. Again be careful of the failed server issuing leases when it comes back online.
2. If you have the option, Windows Server 2008 Enterprise now offers High Availability Clustering for DHCP servers. This is a good solution because the DHCP database is shared between the two DHCP servers, so you don't need to worry about duplicate addresses being issued. This also works across geopgraphically dispersed sites, which is great if you have multiple data centres in different locations. You need to have a storage solution in place thatcan replicate the DHCP database to the failover location.
HTH
Paul

VSOM and VSMS Co-resident in One Server

We have a two CPS-MSP-2RU-K9 Servers with VSOM + VSMS in both the Servers.
That is ( VSOM+VSMS) in Server-1 and again (VSOM+VSMS) in Server-2 .
Can i have Redundancy for VSMS and also for VSOM by this Method ?
Does VSOM in one Server-1 can have Control over VSMS in Server-2 ? Is this a Recommended Design ?

Beginning with VSM7, there are multiple styles of VSMS (media server) redundancy available, depending on the IP Camera matrix you intend to deploy.
There is currently nothing for *native* VSOM redundancy, aside from using VMWare High Availability when deploying VSM OVAs on Cisco UCS platform only.
There is much more that goes into a successful design and deployment than the information you have provided here. My personal recommendation is to leverage your local Cisco ATP partner, or, at a minimum, engage the Cisco PDI Helpdesk (http://www.cisco.com/web/partners/tools/pdihd.html)
Cheers, and best of luck,
Scott Olsen
Solutions Specialist
Bulletproof Solutions Inc.
http://www.bulletproofsi.com

RDS license server failover

Hi
I have a question regarding adding a second RDS license server in my environment. I am using per device mode and want to keep all licenses on my current RDS license server instead of splitting them up. If I add a second server and I loose my primary server
what happens to clients already holding a license and trying to authenticate? Will they be issued a temporary license from the secondary RDS license server or will they be declined since the secondary server is unable to validate the issued license?
Thanks
MeMyselfAndI

Hi,
Thanks for posting here.
For more detailed information about TS Licensing server redundant, you may refer to the following article.
How to provide resiliency to your terminal services licensing infrastructure, focusing on high availability.
http://www.virtualizationadmin.com/articles-tutorials/terminal-services/licensing/terminal-services-license-server-high-availability-recovery-part1.html
Best Regards,
Aiden
Aiden Cao
TechNet Community Support

LAN design - how to implement a core switch?

Hi all,
First post here so please be gentle :-)
I'm looking for a bit of advice with a LAN setup I've been tasked with.
The basic requirements are to have a demonstration suite of servers/storage devices networked with internet access with certain devices segmented in different VLANs. Also, a separate VLAN is required for training and meeting rooms which will receive DHCP addresses from a WIN2K3 server.
The kit I've inherited consists of:
1 ADSL Modem/Router
1 2611XM router
2 Catalyst 4006 switches with Supervisor II engines (CatOS :-( ), one with a layer 3 routing module
Several Catalyst 2950/3500xl switches
Netscreen 100 Firewall
F5 Firepass for VPN
After a lot of fun resetting devices I've currently setup the LAN with a router on a stick configuration which routes between different VLANs (on the 3500/2950s) and which has internet access via the 2611 and ADSL modem router in turn. That's about as far as my current knowledge goes I'm afraid!
What I have to do is incorporate the 4006s but I don't really know how to go about it or what's the best way to use them. How would I use them as core switches?
I was hoping someone could point me in the right direction on the best way to connect the switches up, i.e. network design, cabling (fibre uplinks between switches) and some basic configuration advice with the layer 3 routing module.
Any advice will be most appreciated!! It's my first networking job and I'm a bit lost.
Thanks.

Peter,
I would do the same - with a twist...
Have 1 4006 as a VTP server, also the spanningtree root for all vlans.
Have a trunk between the two 4006's - and make it an etherchannel 2 or 3 ethernet links (redundancy).
Make the second 4006 also a vtp server (redundancy) and have that 4006 the secondary 4006 for spanningtree (more redundancy!)
That way if you decide to have a distribution layer - you have 2 uplinks into the core 1 into 4006-1 as the primary, and the second 4006-2 as the secondary.
You could then have a trunk (etherchannel) between the distribution switches, then have a access layer into the distribution layer with duel links. This way you could have multiple switch and or link failures and still work!!!
You use the layer 3 module to do the inter-vlan routing - correct. Then have your adsl modem/router as the gateway to the internet - you put a default route in the layer 3 module point to the adsl modem! then you have the routes for the various vlan subnets pointing from the modem back to the layer 3 module......done!
HTH.

Server redundancy

Similar Messages

Maybe you are looking for