[cisco-voip] Approximate service outage times for a 5.1->6.1 upgrade

[Darren]

I've done a few of these and here's the way I've done it in the past.

*) If using DNS make sure reverse DNS is working correctly (sub upgrade uses
this regarding the pub)
*) upgrade Pub, reboot to new partition
*) when this comes backup it doesn't become properly active (depending on
config of cluster) so normally the phones will stay with the existing
cluster (again depending on config of cluster).
*) upgrade subscriber and reboot.

Phones will reboot, in my scenario this never took more than 20 minutes,
with 7940's and 60's.  From 200 - 400 handsets.

To do each server takes approximately 2 hours each.  I was under the
impression you had to do it in this order to get the publisher database in a
good condition for the subscriber database to use, but I could be wrong
(anybody able to clarify this).

D




On Wed, Jul 30, 2008 at 11:01 PM, <keli at carocomp.ro> wrote:

> There are close to 1000 phones, a mix of 7940, 7960 and CIPC soft-phones,
> load-balanced across the two servers (so some have the sub as primary, while
> others have the pub as primary in their CCM group)
>
> Also, since the servers are in a datacenter, and the phones in several
> locations in a metropolitan area, I'm not exactly sure of the "distances" of
> each location ...
>
> Anyway, the good news for me seems to be that the 7940/60 phones have a
> much smaller firmware :)
>
> thanks,
>  Zoltan
>
>
> Quoting Kris Seraphine <baryonyx5 at gmail.com>:
>
>  Are all the phones registered with the same box?  If so, your downtime
>> should be limited to the time it takes to upgrade the phone firmware.
>> The SRND has a section on how to calculate this.  I just upgraded 25
>> 7941s to 8.3.5 last night over a T1 line and it took 24 minutes (from
>> start until all phone were registered again).  This was with peer
>> firmware sharing enabled.
>>
>> I'd imagine you'd want to install the license after rebooting the
>> publisher but before rebooting the sub.  The CM service won't start
>> until the license is installed.
>>
>> On Wed, Jul 30, 2008 at 8:49 AM, Kelemen Zoltan <keli at carocomp.ro> wrote:
>>
>>> Hi,
>>>
>>> I'm going to upgrade a cluster of two CUCM 5.1.2 servers (both MCS7845)
>>> from
>>> 5.1.2 to 6.1 release.
>>>
>>> However, the customer is keen on knowing expected service outage times,
>>> and
>>> the duration of the whole operation in itself. Even if they do not expect
>>> exact timeframes, telling them "it will take a lot" just won't do :-)
>>>
>>> So can anyone give me an idea, of how much time would it take to upgrade
>>> this cluster and of this, approximately how much would they be without
>>> service?
>>>
>>> Especially, since, as far as I understand, upgrading in itself does not
>>> mean
>>> interruption is service, since I'm working on an inactive partition.
>>>
>>> Also, is this planned order correct (basically based off SRND):
>>> - upgrade first node, don't reboot
>>> - upgrade second node (there's only one), don't reboot
>>> - reboot first node with switching versions
>>> - reboot second node with switching the version
>>>
>>> the Feature licenses for the v6 should be uploaded before the upgrade
>>> process, or after the first node reboot?
>>>
>>> thanks,
>>>  Zoltan
>>> _______________________________________________
>>> cisco-voip mailing list
>>> cisco-voip at puck.nether.net
>>> https://puck.nether.net/mailman/listinfo/cisco-voip
>>>
>>>
>>
>>
>> --
>> kris seraphine
>>
>>
>
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
> _______________________________________________
> cisco-voip mailing list
> cisco-voip at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-voip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20080730/69202786/attachment.html>