Upgrade from 1.5.0p23 to 1.6.x - questions

General
1

Hi,
I have a stable multi-site setup running 1.5.0p23.
I’d like to test out 1.6.x, but believe I read that a site cannot be downgraded if going to 1.6.x

some questions

  • Any known issues/caveats upgrading to 1.6.0?

  • Would it be better to upgrade from current 1.5.0.p23 to 1.6.0 and then directly to 1.6.0p8? or should I do incremental upgrades between 1.6.0 and 1.6.0p8 ?

2

If you have not installed many interface modifications or inventory plugins the upgrade should be no pain.
Interface modifications and inventory plugins needs some code modifications.
If you want to test the upgrade - make a copy if your main site and try the upgrade with the copied site.

You can upgrade directly to p8. I would only do incremental upgrades if there is a major version in between like upgrade from 1.2.8 to 1.6.

3 Likes
3

Thanks for the suggestion and I did the following with some success:

  • copied site (existing name “network” , cloned name “netwk_stage”)
  • upgraded the cloned site to 1.6.0p8
  • configured the cloned site to use different distributed monitoring port
  • stopped the old site
  • started the new site
  • went into web UI for the new site and adjusted the distributed monitoring connections to correspond with the changes (i.e. each distributed monitoring instance has cloned stage site running 1.6.0P8 as well)

the problem I’m seeing is that I’m somehow seeing 2 entries for what should be the local site when I try to apply changes-- and the weird entry shows “Has never been activated”…

any ideas?
somehow Site ID of the old site is still being read… not sure where I’m supposed to change that?

4

Hi,
what’s the site name in etc/check_mk/multisite.d/sites.mk?

Karl

5

Hello checmk members,

I’ve installed a new server for upgrade testing purpose. Our prod environment is 1.5.0.p21. I restored data from the production server on the test machine, then I upgraded to 1.6.0.p9 version and I started to test some functions.
What I saw after upgrade:
if i delete a host from Hosts it is still persistent in the main dashboard.
also, for some equipments i saw only a empty graph, but in the right side at perf-o-meter column it shows correctly.

Update: both problems were solved after reboot server.

thanks,
Robert

6

Hi Karl,

the master node is the one with weird issue,

entry looks like this

‘network’: {‘url_prefix’: ‘http://REDACTED/network
/’, ‘user_login’: True, ‘insecure’: True, ‘disabled’: False, ‘replication’: None, ‘globals’: {‘wato_hide_varnames’: False, ‘wato_hide_help_in_lists’: False}, ‘proxy’: None, ‘multisiteurl’: ‘’, ‘status_h
ost’: None, ‘socket’: (‘local’, None), ‘disable_wato’: False, ‘alias’: u’San Jose DC’, ‘timeout’: 10, ‘persist’: False},

when I try to edit the site connection in WATO I get this error:

You can only configure a local site connection for the local site. The site IDs (‘network’ and ‘netwk_stage’) are not equal.

the URL looks wrong, it should be http://REDACTED/netwk_stage but for some reason it’s showing up as “network” instead

7

Hi,
you can change “network” to “netwk_stage” and the url_prefix to “/netwk_stage/” in the configuration file sites.mk. Restart the site.

Karl

8

Thanks Karl,

I went back to sites.mk and realized I had to make that change in a few places =
that fixed one issue with the master, but now when I try to activate changes , the slave nodes generate the following error: (all slave nodes)

Internal automation error: Configuration error. You treat us as a remote, but we have an own distributed WATO configuration!
Traceback (most recent call last):
File "/omd/sites/SLAVEXXX_netwk_stage/lib/python/cmk/gui/wato/pages/automation.py", line 186, in _execute_automation_command
html.write(repr(automation.execute(automation.get_request())))
File "/omd/sites/SLAVEXXX_netwk_stage/lib/python/cmk/gui/watolib/sites.py", line 717, in get_request
self._verify_slave_site_config(site_id)
File "/omd/sites/SLAVEXXX_netwk_stage/lib/python/cmk/gui/watolib/automation_commands.py", line 46, in _verify_slave_site_config
_("Configuration error. You treat us as "
MKGeneralException: Configuration error. You treat us as a remote, but we have an own distributed WATO configuration!

any ideas ?

Christian

9

Hello Christian,
have you activated connections on your satellite site? Is WATO deactivated for satellite sites?

Karl

10

Hi Karl,

sorry where do I check for this?
not sure what is meant by activated connections --> all satellite sites were cloned from their existing 1.50p23 counterparts, no change to the clone configuration. is there something I need to turn off?

11

ok found the issue =)
for some reason the satellite sites had the file sites.mk with singular entry for the individual site.
once I deleted this file, things started to work once more =)

3 Likes