High Availability LAN Design

Hi,
Consider the following:
H1------S1----R1----ISP1
H2----- S2----R2----ISP1
H1/2 are dual-homed hosts.
S1/2 are WS-2970G-24's.
R1/2 are 7204VXR's with NPE-G1's connecting to redundant links in providers PoP.
Was hoping to implement HSRP but can this be implemented? The 2970's ( running Enterprise IOS) can't do layer 3 routing, right?
How would you connect redundant links to NPE-G1 gig ports and make them act like two switch ports?
Any advice appreciated.

Apologies. I'll try again.
| H1------S1----R1----ISP1
| \ /\ /
| \ / \ /
| \/ \ /
| /\ \
| / \ / \
| / \ / \
| H2----- S2----R2----ISP1

Similar Messages

High Availability architecture design know-how in 9.3.1

Guys,
Is there any toolkit and know-how how to design high avaialbility architecture (clustering) for system 9.3.1 and particularly HFM.
I believe it could be quite different than in 11th version.
Any input most appreciated
Regards,
Rafal

Hi,
Here is how to create users in Shared Services :- http://download.oracle.com/docs/cd/E10530_01/doc/epm.931/html_cas_help/crtuser.htm
The rest of the information for Shared Services is in :- http://download.oracle.com/docs/cd/E10530_01/doc/epm.931/html_cas_help/frameset.htm?apc.html
Cheers
John
http://john-goodwin.blogspot.com/

Best practice for High availability design, HSRP

Hi,
I am planning to create High Availability for LAN to WAN connectivity.
But I want to know your opinion about the best way how to do this. I googled for a solution/best way how to do this, but I didn't found in my opinion right answer.
The situation:
I have 2 3945E Routers and 2 3560 switches. The design that I am planning to implement is below.
The main goal is to have redundant connection, whatever one of the devices will fail. For example, if the R1 will fail, R2 should become active, if the SW1 will fail, the SW2 will take care about reachability and vice versa. The router 1 should be preferred always, if the link to ISP isn't down, because of greater bandwidth. So why am I drown 2 connections to 2 separate switches. If the SW1 will fail, I will still have a connection to WAN using R1 router.
The Router interface should be configured with sub interfaces (preferred over secondary IP address of interface), because more than 10 subnets will be assigned to the LAN segment. The routers have 4 Gi ports.
HSRP must be enabled on LAN side, because PC's on LAN must have redundant def. getaway.
So, the question is - what is the best and preferred way to do this?
In my opinion, I should use BVI and combine R1 routers 2 interfaces in to logical one and do the same for the R2.
Next, turn the router in to L3 switch using IRB and then configure HSRP.
What would be your preferred way to do this?

Hi Audrius,
I would suggest you to go with HSRP. GLBP you will use where you want load balance.
I think the connectivity between your Routers (3945) and switches (3560) is gigabit connection which is high speed. So keep one physical link from your switches to each router and do HSRP on those router physical interfaces.
In this way you will have high availability like if R1 fails then R2 will take over.
Regarding the config see the below which I have for one of my Customer DC.
ACTIVE:
track 1 interface GigabitEthernet0/0 line-protocol
track 2 interface GigabitEthernet0/0 line-protocol
interface GigabitEthernet0/1
ip address 10.10.10.12 255.255.255.0
ip nat inside
ip virtual-reassembly
duplex full
speed 100
standby use-bia scope interface
standby 0 ip 10.10.10.10
standby 0 priority 110
standby 0 preempt
standby 0 authentication peter2mo
standby 0 track 1 decrement 30
standby 0 track 2 decrement 30
STANDBY:
track 1 interface GigabitEthernet0/0 line-protocol
interface GigabitEthernet0/1
ip address 10.10.10.11 255.255.255.0
ip nat inside
ip virtual-reassembly
duplex full
speed 100
standby use-bia scope interface
standby 0 ip 10.10.10.10
standby 0 priority 90
standby 0 authentication peter2mo
standby 0 track 1 decrement 30
Please rate the helpfull posts.
Regards,
Naidu.

Question on replication/high availability designs

We're currently trying to work out a design for a high-availability system using Oracle 9i release 2. Having gone through some of the Oracle whitepapers, it appears that the ideal architecture involves setting up 2 RAC sites using Dataguard to synchronize the data. However, due to time and financial constraints, we are only allowed to have 2 servers for hosting the databases, which are geographically separate from each other in prevention of natural disasters. Our app servers will use JDBC pools to connect to the databases.
Our goal is to have both databases be the mirror image of each other at any given time, and the database must be working 24/7. We do have a primary and a secondary distinction between the two, so if the primary fails, we would like the secondary database to take over the tasks as needed.
The ability to query existing data is mission critical. The ability to write/update the database is less important, however we do need the secondary to be able to process data input/updates when primary is down for a prolonged period of time, and have the ability to synchronize back with the primary site when it is back up again.
My question now is which replication technology should we try to implement? I've looked into both Oracle Advanced Replication and Dataguard, each seems to have its own advantages and drawbacks:
Replication - can easily switch between the two databases using multimaster implementation, however data recovery/synchronization may be difficult in case of failure, and possibly will lose data (pending implementation). There has been a few posts in this forum that suggested that replication should not really be considered as an option for high availability, why is that?
Dataguard - zero data loss in failover/switchover, however manual intervention is required to initiate failover/switchover. Once the primary site fails over to the standby, the standby becomes the primary until DBA manually goes back in and switch the roles. In Oracle 10g release 2, seems that automatic failover is achieved through the use of an extra observer piece. There does not seem to be anyway to do this in Oracle 9i release 2.
Being new to the implementation of high-availability systems, I am at somewhat of a loss at this point. Both implementations seem to be a possible candidate, but we will need to sacrifice some efforts for both of them also. Would anyone shine some light on this, maybe point out my misconceptions with Advanced Replication and Dataguard, and/or suggest a better architecture/technology to use? Any input is greatly appreciated, thanks in advance.
Sincerely,
Peter Tung

Hi,
It sounds as if you're talking about the DB_TXN_NOSYNC flag, rather than DB_NOSYNC.
You mention that in general, you lose uncommitted transactions on system failure. I think what you mean is that you may lose some committed transactions on system failure. This is correct.
It is also correct that if you use replication you can arrange to have clients have a copy of all committed transactions, so that if the master fails (and enough clients do not fail, of course) then the clients still have the transaction data, even when using DB_TXN_NOSYNC.
This is a very common usage scenario for Berkeley DB replication/HA, used to achieve high throughput. You will want to pay attention to the configured ack policy, group size setting, setting of the 2SITE_STRICT option (if group size == 2).

UCCX 7.0 High Availability IP Addressing

Hi,
I am installing UCCX in HA mode. The servers are on the same site and have a RTT of less than 2 ms.
I am wondering whether to put them in the same VLAN or in separate VLANs. The design guide does not seem to state a preference.
Please let me know what approach works for you

Hello James,
As you mention HA over IP WAN its just support under UCCX 8.0, for now UCCX 7.0 does not support these. That infomation can be check in the SRND.
http://www.cisco.com/en/US/docs/voice_ip_comm/cust_contact/contact_center/crs/express_7_0/design/guide/uccx70srnd.pdf
Page 66 says:
"Cisco Unified CCX high availability requires that the Cisco Unified CCX Engine and Database components and the CTI Managers with which the Cisco Unified CCX servers communicate be located in the same campus LAN and that the maximum round-trip delay between these servers be less than 2 ms"
HTH
Please rate this post if was helpful
Walter Solano
CCVP, Cisco UCCX Specialist
Cisco IP Communications Express Specialist

SharePoint 2013 High Availability - BI Semantic

SharePoint 2013 High Availability
Hi all,
Settings
Two Sites connected via tunnel, no issue in networking – Single farm
SQL2012, two DB with mirroring
SharePoint 2013, two servers in the farm, each server is APP & WEB
Issue, this has been designed against full site failure (e.g. losing connection to other site, which means losing 1 SQL and 1 SP server)
When users try to view some reports from the BI page, they can’t , we have confirmed that SQL mirrors are failing correctly, the issue (we think to be to access the SharePoint reporting services DB which we have mirrored as well), any help will be great.
Regards

Okay, if you've validated that you have the 1Gbps and 1ms over 10 minutes average, then you're in a supported scenario. How did you configure the SSRS databases to use mirroring within SharePoint? With SSRS databases, you'll need to use PowerShell:
$database = Get-SPDatabase | where {$_.Name -eq "ReportingServiceDbName"}
$database.AddFailoverServiceInstance("InstanceName")
$database.Update()
Trevor Seward
Follow or contact me at...
&nbsp&nbsp
This post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

Multi Site SQL High Availability group

Hi,
We have our primary Data Centre where we have 2 MS SQL 2012 Enterprise Servers and 2 SharePoint 2013 Enterprise servers.
We have SQL High Availability group implemented on our primary Site (Site A)
Site A has subnet of 192.168.0.0.
We recently added a new DR Site (Site B). Site B we have MS SQL 2012 Enterprise Servers. Site B have a subnet of
172.1.1.0
Both sites are connected via a VPN Tunnel. MS SQL 2012 Enterprise Server on Site B has been added in have SQL High Availability
group of Site A.
SQL High Availability group have 2 IPs 192.168.0.32 and 172.1.1.32. SQL Listener have 2 IPs 192.168.0.33 and 172.1.1.33
We want to make sure that if Site A completely down then Site B work as an active Site. But when Site A down we are unable
to ping High Availability group and Site B is unable to work as active site. SQL and SharePoint services are completely down.
Site A has AD(Primary Domain Controller) and Site B has ADC(Additional Domain Controller).
Site A has witness server.
We are using Server 2012 Data Centre
Please suggest.
Farooq

SharePoint is not the same as any other applications. The DR site has to be a completely different farm from your production. This means that the SharePoint_AdminContent and config databases on both farms are different and should not be included in the Availability
Group (they do not support asynchronous Availability Group.) Only content databases and other SharePoint databases supported in an Availability Group should be included as per this
TechNet article. Have a look at this
blog post for your networking configuration.
The reason your Windows cluster service goes down in the DR data center when the when your primary data center goes down is because the cluster has no majority votes. You have 4 available votes in your cluster - 3 in the primary data center in the form of
the 2 Availability Group replicas and the witness server and 1 in the DR data center. If the Windows cluster does not have majority of votes, it automatically shuts down the cluster. This is by design. That is why when this situation happens, you need to force
the cluster to start without a quorum as described in
this article (Windows Server 2012 R2 introduced the concept of dynamic witness and dynamic quorum to address this concern.) By default, the IP address of the Availability Group listener name will be registered on all of the DNS servers if the cluster is
running Windows Server 2012 and higher (this is the RegisterAllProvidersIP property that Samir mentioned.) However, you need to flush the DNS cache on the SharePoint web and application servers and point them to the correct IP address on the DR server after
failover. This is because the default TTL value of the Availability Group listener name is 20 minutes.
It is important to define your DR strategies and objectives when designing the architecture to make sure that the solution will meet your goals. This will also help you automate some of the processes in the DR data center when failover or a disaster occurs.
Edwin Sarmiento SQL Server MVP | Microsoft Certified Master
Blog |
Twitter | LinkedIn
SQL Server High Availability and Disaster Recover Deep Dive Course

ASA 5520: Configuring Active/Standby High Availability

Hi,
I am new to Cisco firewalls. We are moving from a different vendor to Cisco ASA 5520s.
I have two ASA 5520s running ASA 8.2(5). I am managing them with ASDM 6.4(5).
I am trying to setup Active/Standby using the High Availability Wizard. I have interfaces on each device setup with just an IP address and subnet mask. Primary is 10.1.70.1/24 and secondary is 10.1.70.2/24. The interfaces are connected to a switch and these interfaces are the only nodes on this switch. When I run the Wizard on the primary, configure for Active/Standby, enter the peer IP of 10.1.70.2 and I get an error message saying that the peer test failed, followed by an error saying ASDM is temporarily unable to connect to the firewall.
I tried this using a crossover cable to connect the interfaces directly with the same result.
Any ideas?
Thanks.
Dan

The command Varun is right.
Since you want to know a little bit more about this stuff, here goes a bit. Every interface will have a secondary IP and a Primary IP where the Active/Standby pair will exchange hello packes. If the hellos are not heard from mate, the the unit is delcare failed.
In case the primary is the one that gets an interface down, it will failover to the other unit, if it is the standby that has the problem, the active unit will declare the other Unit "standby failed). You will know that everything is alright when you do a show failover and the standby pair shows "Standby Ready".
For configuring it, just put a secondary IP on every interface to be monitored (If by any chance you dont have an available secondary IP for one of the interfaces you can avoid monitoring the given interface using the command no "monitor-interface nameif" where the nameif is the name of the interface without the secondary IP.
Then put the commands for failover and stateful link, the stateful link will copy the connections table (among other things) to avoid downtime while passing from One unit to another, This link should have at least the same speed as the regular data interfaces.
You can configure the failover link and the stateful link in just one interface, by just using the same name for the link, remember that this link will have a totally sepparate subnet from the ones already used in firewall.
This is the configuration
failover lan unit primary
failover lan interface failover gig0/3
failover link failover gig0/3
failover interface ip failover 10.1.0.1 255.255.255.0 standby 10.1.0.2
failover lan unit secondary
failover lan interface failover gig0/3
failover link failover gig0/3
failover interface ip failover 10.1.0.1 255.255.255.0 standby 10.1.0.2
Make sure that you can ping each other secondary/primary IP and then put the command
failover first on the primary and then on the secondary.
That would fine.
Let me know if you have further doubts.
Link for reference
http://www.cisco.com/en/US/products/ps6120/products_configuration_example09186a008080dfa7.shtml
Mike

WLC HA, difference between GLOBAL- and AP- High Availability

hello everyone,
I have a question regarding HA and LAP...
we have two 5508 (sw ver 6.0.199.4), on each specific AP we have an entry for which is his primary and secondary controller
so far so good, when one controller fails, the AP is connecting to the second controller and goes on doing his business...
so what I am not sure about is what I should configure globally regarding HA
first question: do I have to configure anything at all?
second question: what should I configure best? we are using our WLCs only to control APs that are connected to our (WLAN-dedicated) LAN, we are not controlling any APs at a remote-location.
finally, let me quote the configuration-guide:
"Follow these steps to configure primary, secondary, and tertiary controllers for a specific access point and to configure primary and secondary backup controllers for all access points."
and the question for this:
what is the difference between a controller and a backup-controller?
from my point of view: if I configure a primary and a secondary controller, the secondary controller is the backup-controller for the primary controller...
while I am writing this, I would like to apologize for what I am asking here, because at this time I am totally confused about this and to write those questions down, did not help to calm down...
thank you very much in advance!
regards,
Manuel

hi Leo,
I tested this out, but i guess its not working as i thought it would work. I configured the backup primary controller IP and name in the global configuration of the Wireless tab of the WLC and left the AP high availability blank with no settings. I joined the AP to the WLC and show capwap client ha output on the AP shows the backup primary controller name. but if i shut down the primary controller, the AP does not join the back, it just tries to get WLC ip by renewing DHCP forever and stuck in that...   below are the outputs.. any idea why its like this ? I thot if there is no HA configured at the AP level, the global config on the controller level should take effect ?
LWAP3-1042#sh cap cli ha
fastHeartbeatTmr(sec)   7 (enabled)
primaryDiscoverTmr(sec) 30
primaryBackupWlcIp      0xA0A700A
primaryBackupWlcName    WLC2-4402-50
secondaryBackupWlcIp    0x0
secondaryBackupWlcName
DHCP renew try count    0
Fwd traffic stats get   0
Fast Heartbeat sent     0
Discovery attempt      0
Backup WLC array:
LWAP3-1042#
*Apr 30 20:36:21.324: %CAPWAP-3-DHCP_RENEW: Could not discover WLC using DHCP IP. Renewing DHCP IP.
Not in Bound state.
*Apr 30 20:36:31.829: %DHCP-6-ADDRESS_ASSIGN: Interface GigabitEthernet0 assigned DHCP address 10.10.114.49, mask 255.255.255.0, hostname LWAP3-1042
*Apr 30 20:37:17.832: %CAPWAP-3-DHCP_RENEW: Could not discover WLC using DHCP IP. Renewing DHCP IP.
Not in Bound state.
*Apr 30 20:37:28.337: %DHCP-6-ADDRESS_ASSIGN: Interface GigabitEthernet0 assigned DHCP address 10.10.114.50, mask 255.255.255.0, hostname LWAP3-1042
*Apr 30 20:38:14.338: %CAPWAP-3-DHCP_RENEW: Could not discover WLC using DHCP IP. Renewing DHCP IP.
Not in Bound state.
*Apr 30 20:38:24.842: %DHCP-6-ADDRESS_ASSIGN: Interface GigabitEthernet0 assigned DHCP address 10.10.114.51, mask 255.255.255.0, hostname LWAP3-1042
regards
Joe

Cisco WLC in High Availability over WAN

Hi my name is Ivan i have a trouble perhaps could you help me...
I have two cisco wlc 5508. I wan to install them in two differents site. One WLC in the site A and the another WLC in the site B.
Site B is the WAN of the site A. The site A is the headquarter.
But i need to configure them in High Availability. For example if the Cisco WLC in site A goes down, the ap's have to registered in the WLC of the site B.
Then the traffic LWAPP have to pass over the WAN between site A to site B.
I have to configure two cisco wlc in HA over a WAN . Please could help me to do this?. Is ok configure the roamming L3 intercontroller?
Thanks for your answers
Regards
Ivan,
AP'S - WLC - SITE A ----WAN-----WLC - SITE B - AP'S
WLC SITE A DOWN = AP'S SITE A REGISTERED IN WLC SITE B

Hi Surendra thanks for yoru answer.
Surendra, if the ap in the site B (in the WAN) goes down then the traffic lwapp have to pass over the wan,
what will should i do to ensure access point can register on to the cisco wlc in the WAN, moreover to configure the mobility groups in both wireless lan controllers?
or i only have to configure in the wlc the mobility groups? Could you explain me what things have i to do to ensure this
SITE A - (ACCESS POINT M) - LWAPP -----PASS OVER WAN---- SITE B - CISCO WLC - (ACCESS POINT M)
STATUS: REGISTERED IN SITE B
Thanks for your answer
IVAN
Regards

High Available SRW2024

I trying to create a high available setup using two SRW 2024 switches and Linux bonded network adapter in mode 1. Each network adapter of the bond is connected to a switch and in case of failure of an adapter or switch, the bonded network adapter will fail over to the other switch using the same MAC address.
I have an aggregated link setup between the two switches. Furthermore I have spanning tree enabled on both switches. If I do a fail-over test and the bond fails over to the other switch, the port state of the aggregated link changes to blocking which blocks all traffic between the two switches.
In case I use a single (none aggregated) link I’m able to prevent the blocking by disabling spanning tree on the inter switch port. In this case the switch needs about 10 seconds to learn the new port of the MAC address. However I’m not able to make this work on an aggregated link. The UI does have the option to disable spanning tree on the aggregated port, but when saving spanning tree is still enabled.
Questions:
1. What is the recommended setup concerning inter switch aggregated links and spanning tree configuration for above mentioned high availability configuration.
2. Why is it not possible to disable spanning tree on aggregated links? Is this by design or is there some problem with the UI?
Any help appreciated.

I have rapid spanning tree enabled on both switches.
The problem is that I have to disabled spanning tree on the link connecting the two switches together. If not the inter switch link will be blocked the moment I fail over the network bond because it probably thinks there is a redundant path. Is there some other way to prevent the inter switch link from blocking?
If not, how can I disable spanning tree on the aggregated link? So far I only managed to do this on a normal link, but cannot do it on an aggregated link.

App-V Streaming in SCCM 2012 - High Availability Possible?

Hello All,
Background:
I'm in the process of designing an SCCM 2012 R2 environment, and one of the things that needs planning for is App-V.
Currently the client uses App-V 4.5 management servers with a mixture of 4.5 and 4.6 apps so I've been told. There are 2 servers and they are load balanced, I think the SQL server used is clustered. I need to bring this App-V environment into SCCM 2012 R2,
and as such I am concerned about High availability. (I'm also aware of compatibility of app v packages but this is beside the point)
SCCM will manage around 7000 workstations, and up to 20,000 users all in all (obviously not at once). The majority of users use any machine, they do not have dedicated ones. The hierarchy will consist of one standalone primary site with SQL on box and I will
be configuring multiple distribution points and management points.
Question:
It is my understanding that Management Points cache policies, and distribution points obviously have all the data cached so to speak. Therefore, if a user is already assigned to a group which is assigned to particular streamed applications - if connection was
lost to the site database server, and they logged onto a new machine, would they still be able to run those streamed applications? Or would the Site database going down break this?
Thanks
Stefan

Yeah that's a good point Torsten, Thanks - this is what I was unclear on and wanted clarification on. Was happy with the answer but thinking about it now you have said that, I dont see how that can work either...
If im getting it right now, it would be ok if someone has already logged onto the machine, as the profile would have already been populated with the virtual application information pulled down after the client had run the evaluation and MP had retreived
the information from the site DB when it was previously online - in that case the application would still work (and be visible) as the client has previoulsy known about it and clicking it can still stream it from the DB.
But in the case of a new login, are we saying that the MP does not cache policies to the extent that it would know that user would be assigned a particular application - it would have to go back to the site DB for this information - which would be unavaliable
and therefore would break it?

ASA 5520 High Availability

I have two ASA 5520s. One has an IDS card and one doesn't. This makes the high availability wizard fail. Can I manually setup the high availability. I don't really need two ASA-SSM-20s. I just want to have one ASA in Standby mode. Is this possible. Anybody have a configuration similar to this?
Thanks,
Alex Pfeil

Hi,
Let me first say that I have not configure a Failover pair where the units dont have matching hardware. So I kind of wonder even if the ASA accepts the configurations, will the failover act normally.
Usually if I configure basic configurations for some ASA Failover pair I first configure all the basic configurations on the ASA and make sure that each interface on the ASA has the "ip address x.x.x.1 255.255.255.0 standby x.x.x.2" configuration under the interfaces
I then configure the existing ASA with a basic Failover configuration such as
failover
failover replication http
failover lan unit primary
failover lan interface failover Management0/0
failover key
failover link failover Management0/0
failover interface ip failover 255.255.255.0 standby
I then prepare a blank ASA and finally configure its Failover too
failover
failover replication http
failover lan unit secondary
failover lan interface failover Management0/0
failover key
failover link failover Management0/0
failover interface ip failover 255.255.255.0 standby
I will then attach the Secondary unit to the network and finally attach the Failover link between the ASAs and let the Primary unit replicate all the configurations to the Secondary unit.
Naturally this is something that would probably be best done during a separate scheduled maintanance break along with configuration backups etc just to be on the safe side.
- Jouni

High Available Attendant console for BE6K?

Hi Arun and team,
Where is it documented whether or not BE6K supports the new release 9.0 High Available Premium Attendant console?
The BE6K data sheet says supports "Cisco Unified Attendant Consoles" and then gives a specification of 500 max users and 12 attendants (implying Business Attendant) and that cant be right since BE6K supports 1000 users.
The Data Sheets for the Attendants dont really say if they are compatible with BE6K.In fact, the CUEAC implies it wouldnt work on BE6K since its says it needs 'specs based support'
http://www.cisco.com/en/US/docs/voice_ip_comm/cucmac/arc/CUACDE_Rel_N_9_0_1_10.pdf
This should be an easy thing to find ;-(
Anyone know where I can find it?
Thanks,
Steve

Nihal
May I suggest kindly cleaning up the documentation so that this is laid out clearly to all of us. This one gives you the impression no change is to be made to the OVA no matter what the console is.
http://docwiki.cisco.com/wiki/Cisco_Business_Edition_6000_OVA/OVF_Templates#Unified_Attendant_Consoles
Then we move to the data sheet
http://www.cisco.com/en/US/prod/collateral/voicesw/ps6788/vcallcon/ps11369/data_sheet_c78-638921.html
Cisco Unified Attendant Consoles is designed specifically for customers with a maximum of 500 IP phones and 12 operators concurrently logged in
So does that mean even if you have premium we can only do 500 Phones and 12 operators? That doesn't sound right.
Next I have this deck from Cisco that says only 25 operators supported. Screen shot attached. So what are the maximums for each type of console supported by the BE6K BU(CUB, CUD, CUE and CUPAC)
Thanks
Srini

Timesten high availability question

I have a case presented here and wanted to know if it is actually possible to implement.
Let us consider four nodes with timesten (11.2) installed in all of them. A datastore with the same name is created on each of the four servers. Two of these servers are in location A and the other two in location B. Servers in location A have replication defined between them and similarly servers in B have replication defined between them. But note that there is no replication defined between any server in location A with any server in location B.
The basic idea of this entire setup is to maintain high availability of timesten at any point of time (in case of natural disasters, etc..) irrespective of the location of the servers
Now, we have oracle software installed in four other systems. Two of these servers are in location A and the other two are in location B. Note that they are not installed on the same box as Timesten.
Scenario 1:
Question: Timesten in location A goes down, how is the high availability taken care of?
Answer: Timesten in the other server in location A should come up and because of the replication process, this will solve the problem.
Is this correct? I think it is.
Scenario 2:
Question: Timesten installed on both the nodes at location A go down, how is high availability taken care of?
Answer: ?
Please remember from above that timesten does not have a replication policy defined between any server in location A with any server in location B. The requirement says that we should be able to recover all the latest data that the nodes at location A had, by pulling it from oracle DB at location A and putting it into TT server in location B. I would like to know if it is possible to do this?

Hello,
Your approach is correct in designing a Disaster Recovery architecture for TimesTen and Oracle Database. TimesTen supports an Active-Standby pair topology that works well with integrating with the Oracle database within a particular site. However, like for any geographical based replication, it is recommended to configure replication across the WAN using the Oracle Database GoldenGate or Streams technologies in ASYNC mode for better throughput and efficiency. It is also recommended to compress replication traffic across the WAN between the Oracle databases.
So while using the Oracle Database to replicate transactions across the WAN is the right thing to do (using Streams Replication or GoldenGate between the two Oracle databases (assuming using an Oracle RAC 2-node cluster in each site), you will not be able to guarantee that any transactions in site A has made it to site B. GoldenGate and Streams technologies have the ability to replicate the data bi-directionally. What this means is that when site A recovers, transactions that had been trapped there (either between TimesTen and the Oracle DB or in the Oracle DB transaction logs), will attempt to replicate again to site B, so it is important to set up a conflict detection/resolution approach which is possible to do in either GoldenGate or Streams.
Note that Oracle Data Guard replication is not supported with TimesTen in such a configuration across the WAN where TimesTen datastores need to be maintained in both sites.
To fully answer the question, however, we should get into the details of the type of cache group tables that you intend to use in TimesTen. If using TimesTen as just a read-only cache while all inserts/updates happen in the Oracle database, then OracleDB would be regarded as the database of record and it would be used to handles all changes while data changes get auto-refreshed from teh Oracle databases in each site into the respective TimesTen tables.
If the application will be looking to take advantage of fast writes into TimesTen using AWT (async writethrough cache group tables), then it is recommended to configure those tables to be DYNAMIC AWT tables so that if a failover to site B takes place and the data is not in TimesTen (but it is in the Oracle Database), it will be automatically loaded on demand as needed from the Oracle database in site B. Note however that there are restrictions that exist with DYNAMIC load on demand cache groups that you need to look into to find out whether those would work in your application's case (particularly, load on demand works only if the where clause includes an equality predicate on the primary keys, foreigh keys, or unique indexes, etc...)
To fully answer your question on non data loss across geographies, you'd have to use Synchronous replication between TimesTen and Oracle using Synchronous Writethrough Cache Groups and SYNC replication in Streams for example between the two geographies. Neither of those configurations are used to my knowledge in the field because they are very non optimal and carry huge response time expense, which slows down replication considerably and affects application response times.
My assumption also is that the need for the Oracle database is because all data does not fit into memory. If the data does fit into memory, then you could also consider a pure timesten replication across the two sites using an active-standby pair on site A and a read-only subscriber on site B that would be made ACTIVE only in the case of a disaster on site A. Once site B takes over, you can also create an active-standby pair in site B based on the newly elected ACTIVE datastore in that site. In all these cases, it is recommended to use SYNC 2-SAFE replication between TimesTen datastores in the same site and ASYNC replication between the two sites.

High Availability LAN Design

Similar Messages

Maybe you are looking for