There is a limit on the size of attachments that are text indexed, as it does not make sense to index huge attachments (their contents aren't meaningful to search), and they can cause memory issues for the application node. The out of box thresholds for indexing different types of attachments are:

In some cases, attachments under the size specified by the threshold can still cause OOM's.

Steps to Reproduce

  1. Node runs out of memory due to "text index events process" or "Regenerating Text Index" is processing large attachment

    2017-02-15 02:59:14 (934) worker.4 worker.4 Starting: text index events process


    2017-02-15 03:00:39 (452) worker.4 worker.4 WARNING *** WARNING *** Exception attempting to extract terms from a xlsx attachment [sys_attachment.sys_id=fb8d1a0c374d3e08b6a7616043990e4c], exception message=[Java heap space] 

    The log message includes the extension type of the attachment: xlsx attachment = Excel
  2. To confirm the extension type of the attachment and it's size:
    • Navigate to/sys_attachment_list.do?sysparm_filter_only=true
    • Search for sys_id'fb8d1a0c374d3e08b6a7616043990e4c' - sys_id from Step 1
    • Download the attachment 


The attachment_extractor table holds information on text indexing properties for each of the different file types. The out of box value for attachment types is either 50MB or there is no value (not suspected to ever be large enough files to cause issues). Depending on memory usage on the application node, this may not be low enough and cause the above out of memory condition.

To provide relief

  1. Navigate to /attachment_extractor_list.do 
  2. Open record for the affected extension type. In this example, it is Extension = xlsx
  3. Set Max Indexable Size (MB) to 30 MB or lower




Related Problem: PRB878976

Seen In

There is no data to report.

Fixed In


Associated Community Threads

There is no data to report.

Article Information

Last Updated:2017-07-10 10:39:03