You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@usergrid.apache.org by "Rod Simpson (JIRA)" <ji...@apache.org> on 2014/02/04 21:30:14 UTC

[jira] [Created] (USERGRID-73) Binary file upload may not be buffering to disk correctly, holding the file in memory

Rod Simpson created USERGRID-73:
-----------------------------------

             Summary: Binary file upload may not be buffering to disk correctly, holding the file in memory
                 Key: USERGRID-73
                 URL: https://issues.apache.org/jira/browse/USERGRID-73
             Project: Usergrid
          Issue Type: Story
          Components: Stack
            Reporter: Rod Simpson


"Anecdotal evidence from production suggests that we are not buffering the file correctly. Scope of this issue should include: 
- verify upload of large files 
- make sure large files are spooled to disk and never occupy more than 5mb of memory, regardless of physical size 
- once written out completely to disk, large files are POSTed to S3 bucket as anticipated 

To generate files of specific size, use the following: 
dd if=/dev/urandom of=/tmp/largefile bs=1m count=20 

This will generate a file exactly 20MB in size from 1MB blocks of random data. 

You will need to include the ""content-type"" attribute on the Asset while testing if you use random data as described above or the content-type sniffer will get confused."



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)