Is there a best practices guide on how to maintain a fairly large Umbraco instance, 11K + nodes? Also, we are in a load balanced environment with the umbraco files, less the APP_DATA/TEMP directory, stored within a DFS. The APP_DATA/TEMP directory is localized for the cache file. I know Umbraco has distributed publishing, but should it be used or should publishing updates be done on each server?
There's two lines of thought here:
1. On a separate server that also hosts the Umbraco sites(db/content) have the users login there, make their updates, and publish the data to the other servers via the built in distributed publishing.
2. Make a copy/staged site of the live site, have the users login and make changes there and then move the changes to the live site either via Courier or another method.
3. Make the changes and updates to the server in #1, but don't use the distributed publishing. Then, on each of the other servers, refresh the cache directly.
We were trying to use Courier to push larger development chunks to the live servers, but, in our experience(Umbraco V4.7.0), Courier kept timing out and wasn't very stable. Are there any other methods other than manually creating the content on the live site from the dev/staging site? Our client wants the ability to make updates throughout the day, so, just running a process at the end of the day is not an option. I haven't had the chance to test the newer version of Courier, so I don't know if the timeouts and instability have been fixed or not - was it Courier or was it the version of Umbraco?
We've also run into stability issues with the cache getting blown out of both the cache file and the cmsContentXml table. We've upgraded to V4.10.11 and applied a patch we found on the Umbraco Development group. The sites are more stable, it appears, at least so far, but we are not really willing to do anything that will start causing timeouts and force the cache to refresh as it takes 30-45 minutes to refresh that table before it creates the cache file - all the while the site is down/unreachable. We are also looking at upgrading, again, to V6.1x, but we would still like to see if there is a recommended method out there to maintain these kinds of sites.
Any feedback would be helpful.
Thanks.
Sean