How do I reduce UWC memory usage when converting large site from Twiki?

Bruce Cran May 31, 2011

I'm attempting to use UWC to convert a large site from Twiki. The problem is it has lots of attachments and while it's going through the stage where it prints "attachment written" the memory usage is rather ridiculous: it appears to want over 24 GB: it pushes my machine which has 12 GB RAM over 12 GB into swap. Is there any way to reduce the amount of memory UWC needs, or do I just need to have plenty of swap space allocated and let it run

4 answers

1 accepted

1 vote
Answer accepted
Laura Kolker
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
June 1, 2011

Hi Bruce,

Reposting the important bits from my answer from the old forum here for doc purposes.

----

So there are two ways to upload attachments:

  1. Xmlrpc (default), which requires 4x as much memory as the size of the attachment
  2. Webdav plugin, which requires less memory, although I can't give you so nicely specific a figure.

If you want to try using that feature, details here: UWC Webdav Attachments. Only downside I've run into is if your attachments have problematic names (like with question marks), then the upload can fail on that attachment.

----

Oh, and one other thought:

If you end up having to increase the memory you've given to your tomcat (which might be a good idea), you probably want to up the memory on the Confluence as opposed to the UWC itself (which you can do as well - let me know if you need instructions for that).

Cheers,

Laura

0 votes
Laura Kolker
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
June 1, 2011

@Tyler re: Sounds like it's reading all the attachments into memory prior to shoving them into the Confluence system.

I could be wrong, but I do not think this is happening. As far as I can see it's just making the xmlrpc call, and it would be doing it one attachment at a time. If anyone wants to take a peak at the relevant code it's here: https://studio.plugins.atlassian.com/source/browse/UWC/devel/src/com/atlassian/uwc/ui/ConverterEngine.java?hb=true#to1366

0 votes
Leandro Nunes May 31, 2011

Hi Bruce,

UWC has a FAQ and one question called my attention:

I have a very large wiki to convert. Are there any gotchas I should look out for?

http://confluence.fangwai.net/display/CONFEXT/UWC+F.A.Q.#UWCF.A.Q.-Ihaveaverylargewikitoconvert.ArethereanygotchasIshouldlookoutfor%3F

Hope this helps you!

0 votes
TylerA
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
May 31, 2011

Sounds like it's reading all the attachments into memory prior to shoving them into the Confluence system. Did you set Confluence/Twiki to store your attachments in the DB or does it store the attachments in the filesystem?

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events