Actually – I think that’s in the steps somewhere. I read that I think.
---
Lelio Fulgenzi, B.A. | Senior Analyst
Computing and Communications Services | University of Guelph
Room 037 Animal Science & Nutrition Bldg | 50 Stone Rd E | Guelph, ON | N1G 2W1
519-824-4120 Ext. 56354 |
Actually, I prefer to disable IMP HA before installing in the inactive PT. This
way, if you have to flip back to the other partition for some reason, IMP is
already in an “HA disabled” state, which tends to make IMP recover a little
better in my experience. Then just enable HA once everything
Personally, I hate the fact that IM can’t be restarted without first
disabling HA. So basically, if you have an unattended restart, you have to go
in and make configuration changes.
Yuck.
---
Lelio Fulgenzi, B.A. | Senior Analyst
Computing and Communications Services | University of Guelph
Yeah, I HATE this bug. Why in the world can’t the docwiki or what ever it’s
called be updated quicker than a bug be filed/made universally known, and who
came up with these recommendations TAC or the BU.
But I have seem IMP just not start services, and adding resources magic fix it.
In the words of the immortal Chris Farley….
holy schnikes
If possible, it is recommended to have 4vCPU and 8 GB RAM as we are seeing more
cases with high CPU due to resources related.
They want 4 vCPU if possible? They think these things grow on trees?
Right now, our two IMP servers are at
That’s the process I typically follow without issues.
Also, I can’t recall if this was posted here, but wanted to make sure you’ve
seen the recent changes to resource requirements for IM This may not apply
to you if you have more than 5000 users.
IM VM resource requirements needs to be
I hear ya. I thought about that. But since we don’t do this very often, and we
do have a window of opportunity, I wanted to take the safe approach.
I would hate to hit an undocumented feature that caused a service outage as
part of upgrading the inactive partition.
---
Lelio Fulgenzi, B.A. |
Lelio,
Are you apposed to upgrading the inactive partition during working hours? For
my planning, the upgrade window starts when the bumps in service occur, which
is usually the switch version.
There are some cases where the upgrade is not “online” there is a R1/R2 upgrade
designation
Sweet. Thanks. I need to map out the window I have and see how this all fits in.
I really hate how the version compatibility is _so_ tight with the IMP cluster
now. One more thing to have to worry about.
Oh well.
TAC just gave me some advice to see how long the DB replication would take
This is what I do for an in-place upgrade with no other changes (hostname, IP
address.. etc).
I do the cucm pub first (obviously), then imp pub (which is more like a cucm
sub for upgrade purposes), then all the cucm and imp subs; everything into the
inactive pt.
I do the pubs individually,
I'm reading the upgrade guide, specifically, the time sequencing:
https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/upgrade/11_5_1/cucm_b_upgrade-guide-cucm-115/cucm_b_upgrade-guide-cucm-115_chapter_010010.html
and it mentions upgrading the IM publisher (to inactive partition) at the
11 matches
Mail list logo
