server consolidation, preserving all historical data

Post general support questions here that do not specifically fall into the Linux or Windows categories.

Moderators: Developers, Moderators

Post Reply
obx2
Posts: 2
Joined: Tue Oct 04, 2005 6:29 pm

server consolidation, preserving all historical data

Post by obx2 »

Hi,

[ sorry folx if this is covered elsewhere. I did search both the docs & the forums, but no luck (keywords used: server consolidation, combining, preserve historical data) ]


We are running cacti in production on multiple Linux servers, each with its own mysql instance (yeah, I know... inherited setup). Until recently, multiple boxes were required due to poller.php+snmpget, and all were still overloaded, with a runtime of about 4min on each server, and a run averages constantly in the 10-15 range.

However, an upgrade to cactid 0.8.6e dropped the run average to ~0.1, and the runtime from 4min+ to 18s. Just beautiful, and stable, too.
(BIG thank you to the developers!!!)

Therefore, we would like to consolidate & decouple the setup, at 3 levels:
(1) single mysql server, with a single cacti database
(2) separate, single poller host for now, multiple pollers later
(3) separate, single web management / graphing engine front-end

Unfortunately, all the current cacti databases appear totally in conflict, with the same keys reused for different devices/variables. I believe I cannot just A) dump db on server1, B) reload on server2, and C) copy rra/*rrd over. (or can I???)

Are there any tools to help with such a server consolidation (must preserve all historical data!) ? Otherwise, would appreciate pointers/sample code/commands to relevant cacti/mysql/rrd internals needed to develop a custom script (will contribute, if so). Can manage scripting just fine, but not sure where to start and how to proceed to ensure no data loss/corruption at any level.

Thank you!!
-- obx
gaurav
Posts: 15
Joined: Fri Aug 30, 2002 8:28 am

Post by gaurav »

I am in the process of doing something similar as well. Unfortunately, there is no simple way.

What I am doing is to create new rrds and graphs for the devices and then copy the existing rrds to the newly created ones. Pain in the royal south but that is the simplest way. Creating of devices and graphs is not very difficult because of templates.
Thanks,
- Gaurav
User avatar
TheWitness
Developer
Posts: 17062
Joined: Tue May 14, 2002 5:08 pm
Location: MI, USA
Contact:

Post by TheWitness »

If I didn't have a full time job already, I would take this as a professional services job to perform the migration. It's not too hard, but difficult. Much of the data in Cacti can be easily recreated without too much effort.

You would prepare the database first, verify it's operation and then rename your rrd's once copied over to the new server. You would perform that process first as a test, and then as a production test once all the bugs were worked out of the process in order to limit any interruption in polling.

TheWitness
True understanding begins only when we realize how little we truly understand...

Life is an adventure, let yours begin with Cacti!

Author of dozens of Cacti plugins and customization's. Advocate of LAMP, MariaDB, IBM Spectrum LSF and the world of batch. Creator of IBM Spectrum RTM, author of quite a bit of unpublished work and most of Cacti's bugs.
_________________
Official Cacti Documentation
GitHub Repository with Supported Plugins
Percona Device Packages (no support)
Interesting Device Packages


For those wondering, I'm still here, but lost in the shadows. Yearning for less bugs. Who want's a Cacti 1.3/2.0? Streams anyone?
Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests