Last Modified: Oct 06, 2020
See more info
Known Affected Versions:
11.2.1, 11.3.0, 11.4.0, 11.4.1
11.5.0, 11.4.1 HF2, 11.4.0 HF4, 11.3.0 HF8, 11.2.1 HF10
Opened: Aug 09, 2013
Related AskF5 Article: K14831
With certain traffic patterns, the hardware compression can exhibit slower processing of compression requests. This can be observed by monitoring the compression stats within BIG-IP system using the command 'tmctl compress' and monitoring the qa-dc provider. The number of requests enqueued, showed in the cur_enqueued stat, is large and drains slowly. The customer may also see watchdog errors reported in the LTM log: Aug 12 10:21:49 B4200-R19-S35 crit tmm7: 01010025:2: Device error: (null) Watchdog on unit 1! The fix for ID425921, ID427607 and ID428150 need to be installed on the customer's platform to resolve hardware compression issues. Note that ID428150 is only required on TMOS v11.2.1 and v11.3.0.
A significant delay can occur before the data is compressed, potentially causing connections to time-out.
This behavior would occur when a large amount of data needs to be compressed at the same time, overwhelming the hardware compression driver.
The fix is to modify the polling behavior in the quickassist driver to allow more efficient handling of hardware compression requests.