Unable to delete Certain Space,files and folders

[predatorz]

Hi All,

I have a problem here (As topic)
I mapped a network drive to alfresco in my workstation, created a folder and was transferring a lot of stuff over to alfresco and my workstation hanged. After which, those folders and files that were transferred over to alfresco, I am not able to remove them away.

I tried from the network mapping to delete them away after halfway but it will fail.

What can i do?

Using Alfresco Community 2.1.0.

[luã_smiguelsequ]

Ok, I managed to recreate the lucene-indexes. It was way harder than I thought.

I needed to copy back the old (corrupted) indexes to alf_data.

Then, with the corrupted lucene-indexes, force dir.root to point to the correct place and index.recovery.mode=FULL — both on tomcat/shared/classes/alfresco/extension/custom-repository.properties and tomcat/webapps/alfresco/WEB-INF/classes/alfresco/repository.properties

Force, on the last file, system.bootstrap.config_check.strict=false

That way, Alfresco will find that the lucene-indexes are corrupt, but instead of aborting because of any errors, it starts the indexing immediately. After half an hour I got the whole index structure back and visible, both on the Web interface and via WebDAV.

Nice, isn't it?

No.

Sadly, every file transfered (either downloading or via WebDAV) will fail with:

The node's content is missing: 
   node: workspace://SpacesStore/90982d6a-0d27-11dc-b594-ddbd2b60ab14 
   reader: ContentAccessor[ contentUrl=store://2007/5/28/16/27/909f595d-0d27-11dc-b594-ddbd2b60ab14.bin, mimetype=image/jpeg, size=0, encoding=UTF-8]  
Please contact your system administrator.‍‍‍‍

However, manually going to alf_data/contentstore/2007/5/28/16/27/909f595d-0d27-11dc-b594-ddbd2b60ab14.bin (in the above example) will find that file there.

So I'm at a tricky point here. The content is all there. Alfresco knows how to relate the node structure to the content. Navigation through directories work fine. It also correctly knows which node points to which file.

It just lost the ability to retrieve the files. Not good. (Also, with errors turned off on the logs, it isn't easy to figure out what is going wrong)

I'm going to try to rebuild the lucene-indexes again, perhaps this time from scratch, and see what Alfresco says this time.

Believe me, in a week, the client will give up on Alfresco. It's not a question of "cost" — they have a budget for getting support if needed — it's just that when Alfresco fails, it fails dramatically and it takes too much time to get it operational again. It's now over 72 hours — 50 of which spent figuring out these issues — since the company lost the ability to use Alfresco. They're not happy.

[steve]

Hi,

I think support is needed.

What version of Alfresco are you using, and is your contentstore (alf_data) on any special filesystem type?

Steve

[luã_smiguelsequ]

Steve,

Thanks for replying back.

Here is the status so far — the system is running on an old Fedora Core 4 server, with an underpowered AMD Athlon™ XP 2400+ and 256 MB, running a standard ext3 filesystem. Alfresco CE 2.0.0. It's being reasonably used by about a hundred users every day (about 10 GBytes of content, averaging 1-100 MBytes per file). Updating the system is out of the question; the configuration is frozen and is to remain like that. Upgrading Alfresco, risking total loss of content, is also not acceptable.

When Alfresco failed, it had an astonishing uptime of 391 days. Not a single document has ever been lost. Although obviously the hardware was underpowered, people in general were happy about the low cost of the server solution and accepted "slowness" as a compromise.

After Java finally gave up its final breath and collapsed dramatically, bringing Alfresco down with it, Alfresco did not manage to start again. So I patiently went through all the options. The last attempt was to bravely delete alf_data (not just the lucene-indexes) and get Alfresco to forcefully rebuild all indexes and reconstruct alf_data from scratch. Alfresco runs again, but only if the strict configuration checking is set to false.

Both on the Web and through WebDAV, all structures and filenames are there. Files can have the metadata saved — and the changes will, indeed, be saved. New content has absolutely no issues. New folders can be created easily; new files can be added both to existing folders, or to new ones. They will be accessible without problems.

All old files (ie. before the crash), however, will never be retrieved. They all fail, apparently without exception, with the dreaded "content is missing" error:

The node's content is missing:
   node: workspace://SpacesStore/c3ca8431-ba19-11dc-b594-ddbd2b60ab14
   reader: ContentAccessor[ contentUrl=store://2008/1/3/17/34/bfdd6b79-ba19-11dc-b594-ddbd2b60ab14.bin, mimetype=application/zip, size=0, encoding=UTF-8] 
Please contact your system administrator.

This happens to any type of file, from simple .txt, to ZIP files, images, whatever. Everything before the crash is like that.

Note that ./alf_data/contentstore/2008/1/3/17/34/bfdd6b79-ba19-11dc-b594-ddbd2b60ab14.bin is there. It's intact. It can be FTP'ed to somewhere else and viewable without any problems.

And also, new items have absolutely no problem. So it's not the filesystem; it's not the paths; it's not the webserver/network (or the newer files would have the same problem); it's not the lucene-indexes (since they get recreated from scratch; I've done it dozens of times, and waited time enough for the alf_data directory to get re-populated); it's not "stale data" on the alf_data structure; it's not even internal database corruption or bad configuration, since Alfresco has absolutely no problem retrieving the new files that get dropped into it after the Big Crash.

Any clues to what I should be looking for are definitely very welcome.

Thanks in advance,

    - Luís

[predatorz]

I did not have any copy of the old lucene-indexes directory.  :cry:
I was able to start alfresco up with an empty sql database, but i do want all the data back, what can i do to achieve this?

Other things that i do to try restore the alfresco.
The full re-index is only 50k+ transactions.
I tar up the whole alfresco directory, untar into a VM, tried to do the full index reovery.
The first few times of re-indexing throws out Java out of memory problem (Java Heap)..
I increase the mem to 1GB, no more out of memory(Yeah!) and while doing re-indexing to 45k+, it throws ERROR [org.hibernate.transaction.JDBCTransaction] JDBC rollback failed. And i do not know what to do again.  :cry:

[predatorz]

Good News!
Using the VM that was created, i increase the tomcat/bin/catalina.sh to 1.5G (JAVA_OPTS="$JAVA_OPTS -Xmx1536m")
and alfresco.sh ('export JAVA_OPTS='-Xms1024m -Xmx1536m -XX:MaxPermSize=1536m -server')
The Index recovery took very long indeed more then a day…
the alfresco index recovery completed 100%.. and i was able to access to the Alfresco once again with the data intact..
After which i copy the lucene-indexes directory back to the old server and it was able to start Alfresco up as well. data seems to be intact as well..

After more 2 months and umpteen times of re-building indexes, i am delighted to get alfresco back again… 
My old server is also running on fc4, pentium 4 1.9Ghz, 512MB RAM.

[steve]

Hi,

Just as a note, you should never edit any files within the Alfresco webapp, you should always use the extension files for changing values such as setting the index recovery mode.

Hi

Unfortunatelly I'm only a business user of alfresco, but according to our sytem administrator you have to write into
tomcat/webapps/alfresco/WEB-INF/classes/alfresco/repository.properties index.recovery.mode=FULL
and restart

sombody please correct me if Im wrong…

hope it helps

daniel

Steve

[steve]

Luis.

Ok, I know what your problem is:

The last attempt was to bravely delete alf_data (not just the lucene-indexes) and get Alfresco to forcefully rebuild all indexes and reconstruct alf_data from scratch

You deleted _all_ your content, not just the index files that can be recovered, but the content of all the files you have ever had uploaded to Alfresco.
That is why you are getting the content missing errors.

You should be able to restore a backup of the content and reindex.

Note, I am very surprised you are managing to run Alfresco with a machine with only 256MB of RAM.
You might want to have a look at configuring the cache values since by default they are set for running in 512MB - take a look here:
http://wiki.alfresco.com/wiki/Repository_Cache_Configuration

Steve

[archana_bonkanp]

Hi All,

We are unable to delete a folder
please help us
when we say delete its taking long time
so we cancelled it in the middle.
the space which we deleted is not viewed in the Left Pane but we are able to see it in the Right pane.
Can you please help us on the same.