Upgrade is going for full indexing always....

[mcasanket]

Hi All,

I am upgrading a system from 3.2 to 3.2.2.7. I also tried to upgrade to 3.4.11 but the same thing happened.
When I start the upgrade server by coping latest alf_data and database even though I have set index.recovery.mode=AUTO it goes for FULL indexing.

If I set index.recovery.mode=NONE or VALIDATE it shows the below messages in log. The dir.root is pointing to correct location even though this error comes. The server does not start.

01:37:43,122  INFO  [repo.admin.ConfigurationChecker] The Alfresco root data directory ('dir.root') is: F:\Alfresco\alf_data
01:37:43,262  ERROR [repo.admin.ConfigurationChecker] CONTENT INTEGRITY ERROR: Indexes not found for 4 stores.
01:37:43,262  INFO  [repo.admin.ConfigurationChecker] You may set 'index.recovery.mode=FULL' if you need to rebuild the indexes.
01:37:43,262  ERROR [repo.admin.ConfigurationChecker] Ensure that the 'dir.root' property is pointing to the correct data location.

The production server never go for full indexing with the same alf_data and database no matter how many times I restart it.

I want to just upgrade the system with 3.4.11 but not want to do full indexing right now. Please suggest.

Thanking you!

[nikes]

Check if lucene indexes folder is empty? If so, it will regenerate index on startup.

[mcasanket]

Hi Nikes,

No the lucene indexes folder is not empty. I copied it from production server itself by getting the server down. I also compered the size of all the folders and they all are same.
The surprising thing is that, no matter how many times I start the production server it never goes for full indexing. But with the same data if I start the new version of alfresco it goes into the full index recovery.

Thank you!

[nikes]

It seems you are using Enterprise Edition. Have you already contacted Alfresco Support?

What you can try on pre-prod environment is,
1. Empty lucene index folder
2. keep recovery mode to auto
3. Empty you tomcat temp and work folder (If you are using Tomcat!)
4. Start the server and let reindexing process finish once.
5. Stop the server and start again.
6. Check if still it goes for re-indexing?

I think it is always good practice to reindex in case of upgrade.

On success, try upgrading as per the suggested upgrade path with custom code compiled with respective version of SDK.

Hope this helps.

-Nikesh

[mcasanket]

Hi Nikes,

Actually we are not able to do the full indexing on 3.2 version as it is taking almost 5 to 6 days to complete just 50% of indexing.
So what I did is, I installed alfresco 3.4.11 on test server and copied alf_data and latest database on this server and then started the server in full indexing.
It took 2 days to do the metadata indexing and 1 day for full text indexing and then the test server was started.
Then I copied these lucene-indexes folder to the production server and started the server and it worked well.

But I am getting  below errors:

Error1:
23:24:40,201 ERROR [pdfbox.filter.FlateFilter] Stop reading corrupt stream
23:24:47,049 ERROR [pdfbox.filter.FlateFilter] Stop reading corrupt stream
23:24:47,049 ERROR [pdfbox.filter.FlateFilter] Stop reading corrupt stream


Error2:
23:25:36,750 ERROR [quartz.core.JobRunShell] Job DEFAULT.ftsIndexerJobDetail threw an unhandled Exception:
org.alfresco.error.AlfrescoRuntimeException: 05070016 Invalid node ref - does not contain forward slash: d8ae998c-0688-4ac5-bf02-54e738c91411
at org.alfresco.service.cmr.repository.NodeRef.<init>(NodeRef.java:91)


Thank You!