Bug ID 707445: Nitrox 3 compression hangs/unable to recover

Last Modified: Jul 13, 2024

Affected Product(s):
BIG-IP GTM, LTM, vCMP(all modules)

Known Affected Versions:
11.6.1, 11.6.1 HF1, 11.6.1 HF2, 11.6.2, 11.6.2 HF1, 11.6.3, 11.6.3.1, 11.6.3.2, 12.0.0, 12.0.0 HF1, 12.0.0 HF2, 12.0.0 HF3, 12.0.0 HF4, 12.1.0, 12.1.0 HF1, 12.1.0 HF2, 12.1.1, 12.1.1 HF1, 12.1.1 HF2, 12.1.2, 12.1.2 HF1, 12.1.2 HF2, 12.1.3, 12.1.3.1, 12.1.3.2, 12.1.3.3, 12.1.3.4, 12.1.3.5

Fixed In:
14.1.0, 13.1.0.8, 12.1.3.6, 11.6.3.3, 11.5.9

Opened: Feb 21, 2018

Severity: 3-Major

Related Article: K47025244

Symptoms

LTM logs show the following message: Nitrox 3, Hang Detected: compression device was reset When the error manifests, there will be three error messages sent to the log over a period of several seconds. The device is then considered unrecoverable and marked down, and will no longer accept compression requests.

Impact

Once the device is marked down, compression will be sent to the software compression provider, until tmm on the device is restarted. This can cause local CPU utilization to climb.

Conditions

This applies only to vCMP guests. Some compression requests can stall the device after a bad compression request is made. Note: Traffic volume and concurrence, along with the type of error have to occur together in order to result in this issue, so the issue is not easily reproduced.

Workaround

There is no complete workaround without a software fix. However, compression will always default to the software compression provider when hardware cannot be recovered. There are three recovery options available if the TMM-internal reset fails to recover the compression device automatically. These should be employed in this order: A. Restart tmm using the command: bigstart restart tmm. B. Restart the vCMP guest. C. Restart the host (which restarts all guests). Note: Because of the traffic volume, timing, and error type that cause this condition, this error might recur. This issue appears to be caused by a particular compression request. So regardless of the recovery method you execute, the problem may recur in a short time, or months later.

Fix Information

Compression device reset recovery made more robust for some compression failures.

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips