Upgrading cucm cluster

Hello everyone. I am facing an issue right now. The last weekend we upgrade our customer cucm cluster.
The cluster has 7 servers but finally we were able to upgrade 5 of the 7. 2 of them failed. I remove the servers from the publisher and turn them off. We need to reinstall the two servers but we do not have the bootable media at wich the cluster finally is working now. I am wondering if it is possible to install a 7.1.3 or another cucm version and after that upgrade to 8.6.2.22900-9.
Does any one has done this.?
Regards

No, all phones. Check out release notes:
CSCub12922 Adding a new subscriber causes device reset
When you add a new node to an existing cluster, all phones that are registered to the cluster are reset
http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/rel_notes/8_5_1/cucm-rel_notes-851.html#wp1957141
HTH, please rate all useful posts!
Chris

Similar Messages

Upgrade CUCM 5.1(3c) to 5.1(3g) and CRS 4.5(2)SR01

I checked the compatibility document for CRS...
http://www.cisco.com/en/US/docs/voice_ip_comm/cust_contact/contact_center/crs/express_compatibility/matrix/crsmatrxeol.pdf
...as we are planning an in version upgrade/patch to our cucm cluster. This document does not list CUCM 5.1(3g) - it only shows up to 5.1(3e). Our current JTAPI version on the CRS server is 3.1(1.4), according to jtapiversion.txt
Will CRS need an update as well in order to continure inter-operating with the new CUCM version?
Thank you

Hi Ram,
This is a supported upgrade path :)
For information about supported Cisco Unified CM upgrades, see the Cisco Unified Communications Manager Software Compatibility Matrix at the following URL:
http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/compat/ccmcompmatr.html
Search Cisco Unified Communications Manager Customer Documentation
http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/search/cscucm.html
System Installation and Upgrade Manual for IP Telephony Enterprise and Midmarket
http://www.cisco.com/en/US/docs/voice_ip_comm/uc_system/UC6.1.1/ipt_system_ins_upg/suimt611.pdf
Cisco Unified Communications System Documentation
http://www.cisco.com/en/US/docs/voice_ip_comm/uc_system/unified/communications/system/ucstart.htm#Cisco_Unified_Communications_System
Cisco Unified Communications Compatibility Tool
http://tools.cisco.com/ITDIT/vtgsca/VTGServlet
Hope this helps!
Rob

Upgrade CUCM Subscriber after the CUCM Publisher, but when ?

Dear Networkers,
Whe we do have a CUCM cluster to upgrade, i know that we should start by upgrading the publisher then upgrading the subscribers one by one.
My question :
I remember i read somewhere that we can start upgrading the subscriber during the upgrade of the publisher but when the publisher reaches a special step in the displayed status in the output of the OS upgrade web page... I can't remember what was that step..
Can you please help ?
Thanks in advance.

Hi Haouaswajih,
Upgrade a cluster:
When you upgrade a cluster to Cisco Unified Communications Manager 9.0(1), begin upgrading the publisher
node first. You can begin upgrading subsequent nodes in parallel after the publisher node reaches a specified
point in the upgrade. The remainder of this section describes how to determine when you can begin to upgrade
the subsequent nodes.
During the upgrade of the publisher node, view the installation log, install_log_.log, by using
the Software Installation/Upgrade window in Cisco Unified Communications Operating System Administration
or the command line interface (CLI). You can begin the upgrade of the subsequent nodes after the following
information displays in the log:
PRODUCT_TARGET is
PRODUCT_NAME is
PRODUCT_VERSION is
You can also use the CLI to search for the relevant information in the install log by following this procedure:
Procedure
Step 1 List the install logs; for example:
file list install install_* date
install_log_2008-10-01.09.41.57.log install_log_2008-10-08.12.59.29.log
install_log_2008-10-14.09.31.06.log
dir count = 0, file count = 3
Step 2 Search the most recent install log for the string PRODUCT_VERSION; for example:
file search install install_log_2008-10-14.09.31.06.log PRODUCT_VERSION
Searching path: /var/log/install/install_log_2008-10-14.09.31.06.log
Searching file: /var/log/install/install_log_2008-10-14.09.31.06.log
10/14/2008 09:52:14 upgrade_os.sh|PRODUCT_VERSION is
7.1.0.39000-97|
Search completed
Step 3 When the file search command finds the PRODUCT_VERSION string in the install log, you can start the
upgrade of the subsequent nodes.
If you want to upgrade the subsequent nodes in parallel with the publisher node, do not choose
the Reboot to upgraded partition on either publisher node or subsequent nodes while configuring
the upgrade options. If selected, the publisher node may complete its upgrade and reboot while
the subsequent nodes are upgrading, which causes the upgrade of the subsequent nodes to fail.
Caution
When you are ready to activate the new version, you must activate the new software on the publisher node
before activating it on all other nodes.
Refer page 23 onwards in the below link,
http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/upgrade/9_0_1/CUCM_BK_U8D523AB_00_upgrade-guide-cucm-90.pdf
Hope this helps.
Anand
Please rate helpful posts !!

Asking for advice for Jabber deployment - multi CUCM cluster\AD domains

I would like some design advice for deploying Jabber and CUPS in our company. We have 2 locations, west coast (SiteA) and east coast (SiteB). Each site have their own CUCM 7.15 clusters, Unity clusters, AD domains (trusted, but not in the same forest).
At SiteA I have setup CUPS (8.6.3.10000-20) and jabber and have it working great.
I would like to setup CUPS\Jabber for SiteB, but they need to be able to IM\call\etc to SiteA (And vice-versa).
SiteA and SiteB both have CUCM LDAP sync turned on, and LDAP directory synced with both domains (although SiteA cannot authenticate to CUCM at SiteB, and vice-versa due to the fact you can only LDAP sync authentication with one domain, CUCM user database contain users from SiteA and SiteB).
We have SIP trucks setup to pass internal calls and line status(BLF) between the trunks, and can communicate via internal extensions just fine.
The problem I’m running into is my jabber-config files uses the EDI directory – which can only look at one domain, so I cannot search the other domain. I believe changing to UDS fixes this, but I understand it would require me to upgrade both CUCM clusters to 8.6.2 - unless I’m mistaken.
I’m aware the desktop sharing will not work until CUCM is upgraded to 8.6.1 or 8.6.2.
I’m wondering if anyone has any advice, or can confirm I’m on the right track. Thanks in advance!

The thing that's important to understand is how CUP and Jabber build the XMPP URI. The URI has a left- and right-hand side; the left is the username while the right is the XMPP domain. CUP uses the LDAP attribute specified in CUCM's LDAP System page, sAMAccountName by default, for the left-hand-side. The right-hand side is the FQDN of the CUP cluster. Jabber must use the same values as CUP when displaying search results. Take note that nowhere in this process does the entire XMPP URI originate from the directory source.
In your case you have two separate CUP clusters in two separate domains. This won't work because when a user searches for a contact in the directory using Jabber, the client will build the XMPP URI as [email protected]. Even if you got the other domain's user objects into the search results the right-hand-side of the URI would be wrong and the presence subscription would never succeed since the other cluster is in another domain. As such your first task must be to move the CUP clusters into the exact same fully-qualified DNS domain. Once this is done you can use Inter-Cluster Peering to build a larger XMPP network in which all users have the same presence domain. If you intend to do Inter-Domain Federation in the future this must be your public DNS domain, not your internal active directory domain. If you use a non-public DNS domain TLS handshake will never succeed for inter-domain federation requests.
Once you have Inter-Cluster Peering in place you can use Active Directory Lightweight Directory Services (the new name for ADAM) to front-end both forests. Both CUCM clusters would need to import the full list of users representing both domains and the sAMAccountNames must be unique across both domains.
Finally, you can instruct Jabber to use UDS and query it's local CUCM cluster which will be able to return a search result from both domains. Since the CUP clusters are peered in the same domain the XMPP URI can be built properly, the presence subscription can be routed to the correct cluster, and life will be good.
By this point hopefully it's clear that EDI won't cut it since it would be limited to only returning search results from the local forest.
Please remember to rate helpful responses and identify helpful or correct answers.

Upgrade cucm 4 to 9.1

Dear All
I have three server (one publisher and two subscribers) , i would like to upgrade CUCM from 4.2 to 9.1 to two UCS 7000 business edition M3 , and i read Jump-flash cut steps to UCS pdf , my Questions :
- to upgrade three servers publisher and two subscribers parallel how it will be ?
- to upgrade from 4 to 9 ,i need to get the backup by dma from 4 to 7.1 then apply or restore UCS ( 7.1 installed on it ) , what is the dma name i need for this step
- instead of going to follow upgrade steps (long story) , can i get backup from version 4 by BARS and prepare it in CSV file then in UCS version 9.1 i can put all phones 1200 phones by BULK file , in other words install immedialtly version 9 on UCS then use BULK file to add phones. it is ok ? any recommendations or concentrations about that ?
I am waiting your updates ASAP
thanks and regards

1. Always upgrade Publisher first and then you can do subs in parallel
2. Data Migration Assistant, download latest one for your 4.2 version
3. You could to some extend, but it would require a lot of manual modifications as the files are different, also there is no export feature in CUCM 4.2 so all entities such as route patterns, CSS, partitions, etc would need to be recreated.
Chris

10.2.0.1 RHEL 4.0 - root.sh CRS - Failed to upgrade Oracle Cluster Registry

I am trying to install 10.2.0.1 CRS on a 2 node cluster on RHEL 4.0 Nahant Update 7. I get the following errors when root.sh is run
Failed to upgrade Oracle Cluster Registry configuration._
Due to this reason , none of the clusterware background processes fail to start, following which the Configuration Assistant screen also experiences failure (Oracle Notification Server Configuration Assistant).
I tried to apply the 10.2.0.4 patchset also, still same issue. Did anyone see this issue ?
1) We use LINUX DEVICE MAPPER for multipathing and udev
2) We configured the multipathing device partitions and bound them to raw devices
So /dev/mapper/oradisk1 bound to /dev/raw/raw1 ( OCR disk)
and /dev/mapper/oradisk2 bound to /dev/raw/raw2 (VD)
I am able to perform the dd command successfully.
Any ideas ?
-Srinivas

BTW, we our OS is Linux x86-64 with RHEL 4.0 Nahant Update 7. I also applied the patch 4679769.
Did a complete uninstall and reinstall. But issue persists.
-Srinivas

Upgrade cucm 7.1 to 9.1

Hi,
From your experience
I'll upgrade cucm 7.1 on MCS to 9.1 on ucs
is there any recommended path that won't face any bugs
Thanks

Hi,
you can refer the link
https://supportforums.cisco.com/discussion/11819286/cm42-86-upgrade?tstart=6990
regds,
aman

Failed to upgrade Oracle Cluster Registry configuration

I'm making a new installation of Oracle RAC environment, the clusterware version is 10 and therefore the RDBMS, I'm using oracle linux 5 update 7, the installation is being performed on only one machine. The problem occurs at the end of the installation when I try to run / u01/app/oracle/product/10.2.0/crs/root.sh the error messages are:
Failed to upgrade Oracle Cluster Registry configuration, which is issued shortly after the execution of root.sh.
And in the log the error message is:
31/10/2011 11:36:36.654: [OCRCONF] [103401184] ocrconfig starts ...
31/10/2011 11:36:36.654: [OCRCONF] [103401184] Upgrading OCR data
31/10/2011 11:36:36.661: [OCRCONF] [103401184] OCR Already in current version.
31/10/2011 11:36:36.669: [OCRCONF] [103401184] Failed to call clsssinit (21)
31/10/2011 11:36:36.670: [OCRCONF] [103401184] Failed to make a backup copy of OCR
31/10/2011 11:36:36.670: [OCRCONF] [103401184] Exiting [status = failed] ...
If anyone has faced this error and won, I'm accepting suggestions.
Att.
Italo Moisés dos Santos.

Hi;
For your issue i suggest close your thread here as changing thread status to answered and move it to Forum Home » Grid Computing » RAC, ASM & Clusterware Installation which you can get more quick response
Regard
Helios

SAP Upgrade in Cluster Environment

Hi,
Need information regarding SAP upgrade in cluster environment(MSCS) like if any prerequisites need to be done.
and if any aditional steps to be performed compared to normal upgrade.
Thanks in advance.
Regards
Ram

the upgrade guide tells you what to do, read it throughly.

How to know firmware version for all devices in a CUCM cluster

Hello,
We need to know the firmware version used for each one of the devices in a CUCM cluster (version 6.1)
Anybody know if we can get this info for example wit the BAT?
Thanks and regards,

You typically need to use a 3rd Party application to do this.
There are many that can do basic device information gathering, but if you are looking for a complete endpoint solution that scales to 10,000+ devices you should have a look at PhoneView from UnifiedFX (http://www.unifiedfx.com)
Thanks
Stephen

Limit of CSR's or UCCE users per CUCM cluster?

Well my math and lack of documentation compete always, so trying to find max number of CSR's one can have, I understand it depends on ucce/ICM environment.
A cucm cluster with 8 subscribers fully loaded hw etc and moderate CSR skill sets and average number of queues:
Some others came up with around 2000 csrs max if we were to poppulate on whole cluster nothig but ucce users or csrs.
Is this limitation because of PG or CTIOS or what? if its due to PG can we have 2 PG's or total of 4 pg's (A and B sides) will this double CSR capacity to 4000?
As you may know capacity of cucm is 30K+ for users and devices.
Thoughts, facts or CCBU advice?
Thanks,
Baseer

If you using CTIOS as your desktop client you can run two PIMs on a CM PG for a max count of 4,000 agents. Unfortunately, if you are using CAD as your desktop client you are limit to a single PIM and I believe an agent count of 1,000 agents. CAD is very dense in relation to CTIOS. Also, if you plan on using a high number of mobile agents you need to count your agents with a heavier load. For CAD, 1 mobile agent translated to 1.7 agent load and CTIOS I believe is 1.3 agent load.
My understanding on the limitiations is with the code in the jtapi interface on the PG. While UCM CTI manager can process a higher agent count the jtapi interface on the PG cannot.
My two cents.

Moving a CUCM cluster

Hi,
I have a few questions, we have a CUCM cluster, one Publisher and two Subscribers and one Voice mail server running on Linux servers from Cisco, version 6.2.
Is it possible to move the servers into a datacenter in another country and run the servers on VMware?
Can you install survilianse applications on a Cisco Linux CUCM server, like IBM's TSM?
Have anyone done this before so there is some examples on successfull moves?
Regards
Stellan Sveman

Brent,
I dont know if you implementing UC on UCS B or C series servers, but here are some links that have helped me in the past:
UC on UCS CUCM 8 (UCS C-Series foucused)
http://www.youtube.com/watch?v=6_WPGjtZCKY
Cisco Unified Communications on the Cisco Unified Computing System
http://www.cisco.com/en/US/prod/collateral/voicesw/ps6790/ps5748/ps378/solution_overview_c22-597556.html
Unified Communications Virtualization Sizing Guidelines
http://docwiki.cisco.com/wiki/Unified_Communications_Virtualization_Sizing_Guidelines
I hope these links help, and I apologize if you already have the above information.

CUCM Cluster "Partition Unalligned" Issue Resolution by PCD (Prime Collaboration Deployment)

Dear all,
I will test soon (asap i got bootable 9.1.2 software) to solve "partition unalligned" issue with CUCM Cluster version 9.1.2.11900-12.
In my opinion, we can solve this issue with PCD as migrate unalligned cluster to a new cluster with same version.
Do anybody test it before?
What is your opinions?
Best Regards,
Mesut

Hi Aman,
As i found out, correct answer should be "migration to 9.x unsupported by CPD" for CUCM version 9.X.
But just imagine, you have 8 servers in a cluster at version 10.X with Partition Unalligned issue. Is PCD not a good alternative to DRS to solve this issue while migrating to a new cluster with PCD?
Best Regards,
Mesut

Moving CUCM Cluster from 7.1.5 to 10.5 and migrating P2V

Community - What I am looking for is advice from those who have actually done it! Reading about sex and having sex are two different subjects. I am finding relevant docs on the CISCO site, some are contradictory. One doc on CISCO Prime Deployment says yes to CUCM 7.1.5 and another says no. We know the standard "refresh" upgrade process, cop files and that entire drill. What we are looking for is comments from those who have done it before! Our Cluster is mixed MCS servers with CUCM PUB and SUB at Version 7.1.5. Unity/UCCX at 8.0.2 with no HA (don't ask). We want to end up with BE6000 and have UCCX in HA mode on Version 10.5 (10.6)! So what do you think, Migrate all in cluster to a common base like 9, then migrate to VM and upgrade? Or try to do it some other way.
All Comments welcome and very appreciated!

@Jamie I think we eliminated PCD as a strategy. We see this as following the usual upgrade process and assume we will be taking multiple steps. Step one would be to get all on a common version and we are thinking 8.5. At that point we could do the P2V and then upgrade to 10.5 on ESXi. We need to create some HA for Unity and UCCX, but will do that after we get to 10.5. Hopefully we can upgrade the CUCM Publisher, during which time the rest of the cluster will be operational? Put Pub back online as 8.5 and then do the subscriber. Again, we think UCCX and Unity will be operational on PUB as we upgrade SUB. Once on CUCM 8.5 we will then schedule maintenace windows to upgrade Unity and UCCX. At that point we have a cluster on 8.5 ready for P2V. We do that migration and when all are on ESXi, we do the upgrade to 10.5 and add in a HA server for UCCX and Unity. How does that sound? As we really dont want phones rebooting and downloading firmware, we will do that ahead of time, but are researching the level we should get them to before the 8.5 upgrade.
@Amn thanks for the link. We have been reading all we can on the CISCO site, and are really seeking out those who have done this and can share experience.
Thanks to all who comment.

Upgrade cucm from 8.5 to 9.1 : Security issue (CTL Files)

Hello All,
I am in the process of performing a P2V upgrade from CUCM 8.5 to CUCM 9.1. The same IP address and Hostname on the new cluster will be kept.
Could you please advise on what are the right steps to do this operation, knowing that I have deployed CTL files on my phones using usb etokens.
My cluster includes 1 PUB + 8 SUB distributed over WAN.
Total phones is : 1800 phones
Only the Publisher will be upgraded to virtual instance. All subscribers servers will be kept at mcs.
Actually, I'm working on the Publisher to restore the cucm DB v8.5, do an upgrade to v9 and test all phones behavior before upgrade the whole cluster.
Do i to take care about subscribers before upgrade, knowing that tftp role is on different servers not only the publisher ?
What is the behaviour of phones when going to the new cluster ?
What is needed to be done in maintenance mode and in production mode ?
Thanks,
Driss

What version of 8.5.1 are you on? Look out for https://tools.cisco.com/bugsearch/bug/CSCtn50405/ which causes certificates to not be backed up properly so when you do the backup/restore you could run into problems. The good news is that you are using CTLs which uses a physical token so you cannot run into a state where the phones do not trust anything (unless the tokens are lost/damaged of course).

Upgrading cucm cluster

Similar Messages

Maybe you are looking for