We are experiencing a strange problem with a very large Umbraco installation that we just upgraded to v. 6.0.0.
Just after the upgrade, everything seemed to be working correctly, but suddenly almost all of the nodes lost their publish-status. The nodes are published, but when I go into the properties tab the "Link to document" just shows this text: "Oops: this document is published but is not in the cache (internal error)".
That means that the website is not working at all because it seems like the nodes are not in the cache.
I tried to "Republish entire site", I tried to delete umbraco.config and touch web.config to make Umbraco rebuild the cache, but it doesn't help.
When I manually publish the nodes, it works again, but then suddenly they get to the same state again.
Anyone tried this, or have an idea of what the problem could be?
What was you upgrade process? And from what version did you upgrade?
Perhaps the issues you're seeing is because of some missed upgrade steps?
If not I urge you to file a bug report with the above description + info on the version you came from at the issue tracker here http://issues.umbraco.org/issues/U4
But let's first try to make sure that this is not related to missed steps during the upgrade.
To eliminate that the error is caused by some of our own code, we have now tried to comment out the subscriptions to the AfterPublish Events where we made some calls to update the cache. Apparently this has helped, but I cannot tell for sure yet, because the symptoms we saw were intermittent.
I will monitor the site for the next couple of days, and I will return with a conclusion afterwards.
Unfortunately it didn't help, as I witnessed the error occurring again yesterday.
Looking at the Umbraco Log, I found out that it always starts with this entry "[UmbracoExamine] Adding examine event handlers for index providers: 3", followed by thousands of "Save and Publish performed by user" (about 3-5 every second). And the userId in the log is the user that created the node initially.
Is this the normal behaviour of the examine indexing?
By the way - when I experience the problem now, I can republish entire site, and the nodes are all working as expected.
We have managed to locate the issue. There is a timer that triggers publishingService.CheckPublishing() with some interval. Here Document.GetDocumentsForRelease() is called to determine what documents should be published based on a release date. Unfortunately, this query will also return documents, where an older version exists with a release date (i.e. rows in the table cmsDocument where newest = 0 and releaseDate != NULL). This must be a bug in the query.
As a result Umbraco, will then endlessly re-publish the nodes over and over again. This will make the database grow in size extremely fast and ultimately make umbraco fail.
We are not sure how it relates to the original issue "Oops: this document is published but is not in the cache (internal error)".
Strange publishing problems after upgrade to 6.0.0
Hi,
We are experiencing a strange problem with a very large Umbraco installation that we just upgraded to v. 6.0.0.
Just after the upgrade, everything seemed to be working correctly, but suddenly almost all of the nodes lost their publish-status. The nodes are published, but when I go into the properties tab the "Link to document" just shows this text: "Oops: this document is published but is not in the cache (internal error)".
That means that the website is not working at all because it seems like the nodes are not in the cache.
I tried to "Republish entire site", I tried to delete umbraco.config and touch web.config to make Umbraco rebuild the cache, but it doesn't help.
When I manually publish the nodes, it works again, but then suddenly they get to the same state again.
Anyone tried this, or have an idea of what the problem could be?
Hi Rasmus
What was you upgrade process? And from what version did you upgrade?
Perhaps the issues you're seeing is because of some missed upgrade steps?
If not I urge you to file a bug report with the above description + info on the version you came from at the issue tracker here http://issues.umbraco.org/issues/U4
But let's first try to make sure that this is not related to missed steps during the upgrade.
/Jan
Hi Jan,
We upgraded from 4.10.1 to 6.0.0 following the steps from http://our.umbraco.org/documentation/Installation/Upgrading/general very carefully.
To eliminate that the error is caused by some of our own code, we have now tried to comment out the subscriptions to the AfterPublish Events where we made some calls to update the cache. Apparently this has helped, but I cannot tell for sure yet, because the symptoms we saw were intermittent.
I will monitor the site for the next couple of days, and I will return with a conclusion afterwards.
Thanks for your reply.
Rasmus
Unfortunately it didn't help, as I witnessed the error occurring again yesterday.
Looking at the Umbraco Log, I found out that it always starts with this entry "[UmbracoExamine] Adding examine event handlers for index providers: 3", followed by thousands of "Save and Publish performed by user" (about 3-5 every second). And the userId in the log is the user that created the node initially.
Is this the normal behaviour of the examine indexing?
By the way - when I experience the problem now, I can republish entire site, and the nodes are all working as expected.
Rasmus
We have managed to locate the issue. There is a timer that triggers publishingService.CheckPublishing() with some interval. Here Document.GetDocumentsForRelease() is called to determine what documents should be published based on a release date. Unfortunately, this query will also return documents, where an older version exists with a release date (i.e. rows in the table cmsDocument where newest = 0 and releaseDate != NULL). This must be a bug in the query.
As a result Umbraco, will then endlessly re-publish the nodes over and over again. This will make the database grow in size extremely fast and ultimately make umbraco fail.
We are not sure how it relates to the original issue "Oops: this document is published but is not in the cache (internal error)".
A bug has been reported: http://issues.umbraco.org/issue/U4-1783
is working on a reply...