Hi,
we have three large corporate intranets on 3.18 that we are planning to go live with in the next couple of weeks. The largest of the three has around 7,500 media assets and 3,800 pages, and that's the one giving us grief at the moment.
I tried to set the site root live and apply public read ( cascading ), but after a massive hipo task and three hours I got some small percentage of assets updates, and a mass of errors. I subsequently resorted to tackling it by subsections… which also chucked heaps of errors on the more content heavy ones ( smaller ones were OK though ). Drilling down to lower levels I now find maybe one third of the assets have had permissions applied and gone live, the others haven't - and these are scattered all over the site. It's a mess.
Given the complexity of the sites and levels of navigation, it would take weeks to laboriously go through and manually set permissions and make live every asset the numerous hipo tasks randomly missed. But un applying / re-applying at a higher level is sure to meet with the same fate as earlier attempts.
I am wondering if there is any solution to this very annoying problem? How exactly do you make a large site live when the hipo tasks repeatedly throw masses of errors? We have three different installs of Matrix on different boxes and I've seen all of them do this at various times. This one is a brand new install, only been up and running for three weeks.
Wondering if there is even a system tool that will apply permissions / live status to a root node? It would be pretty handy right now!
There are a few things you can do.
First is to try Squiz Server. If that is bombing because it runs out of memory, you'll need to run without Squiz Server but with a high status change threshold (like 50). This will take longer, but complete more reliably. it may even be faster in the end for really large sites.
What else might be happening is that your locks are expiring. In version 3.18+, we no longer keep the locks for so long so you don't have any problem with locking. For lower version, you should increase the lock length on the System Config screen from 10 mins to something really big like 10 hours. Once the HIPO is done, change it back. This will ensure your locks are not lost during the status change.
Also, you can re-apply the live status to all assets that were missed by starting at the top of the tree again and clicking the Cascade checkbox next to the status drop-down. Even though you are not changing the status, Matrix will try and re-apply that status to all children.
[quote]There are a few things you can do.
First is to try Squiz Server. If that is bombing because it runs out of memory, you'll need to run without Squiz Server but with a high status change threshold (like 50). This will take longer, but complete more reliably. it may even be faster in the end for really large sites.
What else might be happening is that your locks are expiring. In version 3.18+, we no longer keep the locks for so long so you don't have any problem with locking. For lower version, you should increase the lock length on the System Config screen from 10 mins to something really big like 10 hours. Once the HIPO is done, change it back. This will ensure your locks are not lost during the status change.
Also, you can re-apply the live status to all assets that were missed by starting at the top of the tree again and clicking the Cascade checkbox next to the status drop-down. Even though you are not changing the status, Matrix will try and re-apply that status to all children.[/quote]
Thanks Greg, 10 hours did the trick! 