November Happy Hour will be moved to Thursday December 5th.
AI OnAI Off
November Happy Hour will be moved to Thursday December 5th.
When PTB updates page types it will trigger a "clear cache" event, this event going to be propagated to all other machines with EPiServers remote events when the site starts up. So what you probably can do is; when you disable PTB update also disable remote events and enable it when the PTB updating is set to off.
What we are doing is to enable PTB auto update for the very last server in deployment list.
But you have to disable remote events on that machine when it's restarted with PTB updating on, otherwise it will send out a "clear cache" remote event to all other machines – and all caches will be cleared out on all other machines.
I have came up with following PS script that disables Events and then re-enables it afterwards.
$fullPath = "C:\Dev\CMS\Trunk\src\CMS.Web\Web.config"
[xml] $xmlOriginal = Get-Content $fullPath
$xmlDisabled = $xmlOriginal.Clone()
$serviceModelXML = $xmlDisabled | Select-Xml -XPath "configuration/system.serviceModel"
$xmlDisabled.configuration.RemoveChild($serviceModelXML.node)
$xmlDisabled.Save($fullPath)
## SPAWN UP BACKEND HERE TO UPDATE PTB
## ONCE SPAWNED UP RUN FOLLOWING TO REVERT CHANGES TO WEB.CONFIG
## WAIT FOR DEPLOYER TO PRESS ENTER TO CONFIRM BACKEND IS UP AND PTB UPDATED ALL PAGES
$xmlOriginal.Save($fullPath)
Hi all,
We have currently 4 public facing servers and one thats used only for backend content updates. Our current deployment below;
1. Deploy backend with page type builder update on (to get all properties updated)
2. Disable backend (app pool disabled)
3. Then deploy each of the public facing servers (it only reads so no PTB etc)
4. Re-enable backend once all sites are up and running.
Issue we are having now is that when we deploy backend with PTB updates for some reasons all public facing websites go down. There is nothing in EventLogger, log4net shows Request Time Out. Performance monitor shows Cache Total Entries drop to 0. This causes our sites to fully go down despite being in load balanced environment.
I'm not exactly sure how to debug it. My gut tells me it could be multicast UDP packets telling all servers that content has updated and therefore causing them to drop cache. Could it be caused by PTB? Has anyone had any issues of similar nature?
Thanks,
Luke