Urgent help needed. Can't rebuild internal index. ...
# help-with-umbraco
m
For months I have been chasing an issue on my site where the internal index was dropping content nodes from the index and I could not rebuild it. It was confirmed by Umbraco Support that this was a known bug. They told me till a fix is released that I should just rebuild the index. The problem is that if I hit the "Rebuild index" I just get the message "The process is taking longer than expected, check the Umbraco log to see if there have been any errors during this operation". If I set he log level to "debug" no errors ever show up in the log but the internal index never gets populated. I have posted on GitHub and was told to update the Examine version, which did not fix it, then they told me to contact Umbraco Support who (despite this only happing in my cloud hosted environment with them) told me this is not something they can support, even though they did also at one point tell me it was a known bug.
b
When you rebuild your index, go to your Settings then Logs and what kind of error you get?
m
That is the thing the logs there don't ahve any errors. Jsut two entires. "Rebuilding index '"InternalIndex"'" and "Starting async background thread for rebuilding index "InternalIndex"." after taht nonthing. This seems to be a unique thing to Azure hosted sites. I did just get a message from cloud support they are going to check more on there side to see where they should be supporting it at. I was also exploring the idea of moving off UmbCloud to another provider and after talking with them it seems like with there setup this would not be a problem and might actually be cheeper for me to move to them. I work night shifts and am very tired right now as I am about 2 hours past when I normaly go to sleep so I am disconnecting and wil lbe back in like 8 hours to continue tackling this from all possible angles
So doing some talking to a few people and digging this is an issue in the cms with azure hosting setup, which cloud runs on. I am now investiganting a new non azure based host as I can see (and would agree) this is not a cloud platform issue but a cms issue so the likely hood it is fixed in a time frame that does not significantly impact me is small.
d
hi Matthew, did you try to delete the current index files first?
m
Yes. If I delete the index and restart the environment it does rebuild the index but only partially. The external will have like 1600 items in it and be correct but the internal will have about 1200 and I can easly find many missing documents. If I hit rebuild after that it gives me the taking longer message and no errors loged.
m
Might be one to try if you are using blockgrid/list but not necessarily the nested props.. and reduce the index size.. https://docs.umbraco.com/umbraco-cms/v/10.latest-lts/reference/configuration/indexingsettings
m
I do use block lists that have RTE fileds, is this considered a "property" in the block? If so I do need that info indexed
With the help of @AaronSadlerUK I was got setup on there hosting and ended up replicating the issue but this time the issue was throwing errors to the log and they helped me determain that it appears to be a package installed that is failing on a JSON token cast during the indexing of complex properites. Cannot cast Newtonsoft.Json.Linq.JObject to Newtonsoft.Json.Linq.JToken. The package is the Limbo.Umbraco.Tables. This package was working with no problems for the longest time then seems to have been broken during an install of one of the Umbraco 13 minor patches. So something Umbraco changed in a patch seems to have broken the package for me. I don't know why I could not get errors to log when in Umbraco Cloud but I now know the root cause and have opened a issue with the package creator.
t
This can happen when you go from one environemnt to another and bring over the Umbraco Data Cache folders. Umbraco - Data - Temp _ examine indexes - Delete everything int here. Then try and rebuild the indexes. If not then try deleting the mainDom file and everything in Nucache
m
The thing was that even deleting the indexes was not helping. Have no idea why errors where nit getting loged but it looks to have been figured out and i opened a bug repoet with the dev. I actually plan on downgrading my umbraco version and slowly upgrading and resting to see which patch broke the package.
a
I have submitted a PR which fixes this issue. https://github.com/limbo-works/Limbo.Umbraco.Tables/pull/40 cc @Anders Bjerner
m
I really appreciate the help you providing in figure out the root cause of my indexing issue as well as submitting a fix for it. You have saved me a lot of time, aggravation, and money.
d
+1 for @AaronSadlerUK being super helpful fixing DevOps issues. I always recommend him.
148 Views