Poor performance importing data

[sbenedetti]

I'm using Alfresco Community 4.2d and I'm trying to import files from a network share.

I'm importing for example 3 or 4 GB of data and during import CPUs reach 100% and Alfresco become unresponsive.
CPUs go high and stay 100% for a lot of time (from half to an hour) and then they come back to a normal load.

I've tried to import data using both CIFS and the builtin BulkImport tool but the problem is always the same.

My configuration:
- installed Alfresco out of the box onto ubuntu desktop 12.04 64 bit
- postgresql
- virtual machine with 4gb ram and 4 xeon processor
- CIFS on

I've tried to switch off multithreaded tracking for SOLR with no luck.

Any help?

Thank you

[mrogers]

Your VM probably does not have enough RAM.

Content ingestion is always a heavy operation but half an hour does not seem right.   I suspect your JVM is doing something sub-optimal like swapping memory.    You need to do a little profiling to see where the bottleneck is.   

Also remember to gather stats on your database after doing a big ingestion.

[sbenedetti]

Virtual machine has 4gb of ram.

Or are you talking about JVM configuration?

Thank you

[mrogers]

The JVM needs to fit without paging into your VM.

[sbenedetti]

How can I see running JVM values?
And how can I set them?

[mrogers]

jmx. Use something like jconsole.   Or a profiling tool.

[sbenedetti]

Ok I take a look at JConsole

And to set values instead? Which configuration file or script I have to edit?

Thank you very much for your support!

[sbenedetti]

I've seen this during tracing:

java process shows a futex_wait_queue_me

What is it?