Nutanix Alert - A1068 - FingerprintingDisabled

Nutanix Alert - A1068 - FingerprintingDisabled

Nutanix Alert - A1068 - FingerprintingDisabled

Description

This Nutanix article provides the information required for troubleshooting the alert FingerprintingDisabled for your Nutanix cluster. For an overview of alerts, including who is contacted when an alert case is raised, see KB 1959.

Note: You can review the specific clusters affected by this alert via the discoveries on the Support Portal powered by Nutanix Insights.

Alert Overview
The FingerprintingDisabled alert is generated if deduplication is automatically disabled when metadata exceeds 250 GB or 50% of the default threshold of the metadata usage on any one of the Controller VMs in the cluster.

For more information on fingerprinting (deduplication) refer to Prism Web Guide 

Sample Alert:

Block Serial Number: 14SMXXXX0000
alert_time: Sat Oct 22 2016 15:31:23 GMT-0700 (PDT) 
alert_type: FingerprintingDisabled 
alert_msg: A1068:Disabling fingerprinting (deduplication) for future writes.
Calculated metadata usage of 128928090328 bytes exceeds safety limit of 128849018880 bytes on Controller VM 27886XXXX. The metadata disk size is 638580662272. 
cluster_uuid: 000xxxxx-xxxx-xxxx-0000-00000000xxx
cluster_id: xxxx 
alert_body: No Alert Body Available

Output messaging

Check ID
130021
Description
Fingerprinting Disabled
Causes of failure
The metadata disk is used for metadata and data. Once the metadata portion on the disk exceeds the safety threshold AOS disables fingerprinting (deduplication) of future writes. Metadata sizes can increase due to a variety of reasons such as having too many snapshots, or simply because too much data is fingerprinted.
Resolutions
Once the metadata size is brought under the safety limit (for example, by removing unneeded snapshots, or by expanding the cluster size), the fingerprinting will be auto-enabled.
Impact
The advantages of deduplication are not available.
Alert ID
A1068
Alert Title
Fingerprinting Disabled
Alert Message
Disabling fingerprinting (deduplication) for future writes. Calculated metadata usage of metadata_usage bytes exceeds the safety limit of metadata_limit bytes on Controller VM service_vm_id. The metadata disk size is metadata_disk_size.

Solution

Note: Due to the SSH limitations for NC2 on Azure clusters, follow this guide to enable the remote support tunnel and reach out to Nutanix Support in order to resolve this issue.

Note: There is a known deficiency that causes a false alert to be generated even if no storage containers use deduplication. This specifically affects AOS versions 5.20.2 to 5.20.4 and 6.0.2 to 6.1.2. This false positive occurs when the metadata usage on the cluster is high enough to need to temporarily disable fingerprinting until the garbage collection systems in place catch up. After that reenable it. In this case, no fingerprinting occurs and there is no concern to address.

Troubleshooting

1.  Verify whether deduplication is enabled on any of the storage containers in a cluster. If no containers have dedup enabled, you can safely ignore this alert as false-positive.

nutanix@cvm:~$ ncli ctr ls | grep -E "  Id| Name | Finger| On-Disk"
 Id : 00057ee3-76e3-7134-6547-ac1f6b5b7075::10
 Name : NutanixManagementShare
 Fingerprint On Write : off
 On-Disk Dedup : off
 Id : 00057ee3-76e3-7134-6547-ac1f6b5b7075::492
 Name : SelfServiceContainer
 Fingerprint On Write : off
 On-Disk Dedup : off
 Id : 00057ee3-76e3-7134-6547-ac1f6b5b7075::9
 Name : default-container-91507931805487
 Fingerprint On Write : off
 On-Disk Dedup : off

2. Verify whether Deduplication is disabled on the cluster by running the following command on any CVM:

nutanix@cvm$ ncc health_checks stargate_checks dedup_auto_disabled_check
Running /health_checks/stargate_checks/dedup_auto_disabled_check            [ PASS ]
-----------------------------------------------------------------------------------+
+---------------+
| State | Count |
+---------------+
| Pass  |   1   |
| Total |   1   |
+---------------+

If the value returned is PASS, notify Nutanix Support that Fingerprinting has been re-enabled. Otherwise, if WARN or INFO, continue to Resolving the issue section.

INFO: Fingerprinting/Dedup is disabled by stargate.
WARN: Fingerprinting is disabled by stargate, but configured on Storage Container(s): [u'xxx']

Resolving the Issue

If NCC confirms that Fingerprinting is disabled on the cluster, consider engaging Nutanix Support. Collect additional information and attach to the support case.

Preventing the issue
Here are some recommended measures you can take to try to avoid this issue:

  1. Make sure the Cassandra metadata ring is balanced. See KB 1342 for details.
  2. Upgrade SSDs to higher capacity drives.
  3. Avoid having clusters of nodes with different capacities. This issue can be caused by different SSD tier sizing between the nodes.
  4. Expand the existing cluster.
  5. Make sure cluster workload is suitable to benefit from deduplication. Refer to Deduplication Best Practices section in the Prism Web Console Guide.

Note: This might be different for each Storage container. Only enable deduplication where appropriate. 

Collecting Additional Information

  • Before collecting additional information, upgrade NCC. For information on upgrading NCC, see KB 2871.
  • Collect the NCC output file ncc-output-latest.log. For information on collecting the output file, see KB 2871.
  • Collect the Logbay bundle using the following command. For more information on Logbay, see KB 6691.
    nutanix@cvm$ logbay collect --aggregate=true


Attaching files to the case

When viewing the support case on the support portal, use the Reply option and upload the files from there.
If the size of the NCC log bundle being uploaded is greater than 5 GB, Nutanix recommends using the Nutanix FTP server due to supported size limitations. See KB 1294.

Requesting assistance
If you need further assistance from Nutanix Support, add a comment to the case on the support portal asking Nutanix Support to contact you. If you need urgent assistance, contact the Support Team by calling one of our Global Support Phone Numbers. You can also click the Escalate button in the case and explain the urgency in the comment, and Nutanix Support will be in contact.

Closing the case
If this KB resolves your issue, and you want to close the case, click on the Thumbs Up icon next to the KB Number in the initial case email. This informs Nutanix Support to proceed with closing the case. You can also update the support case saying it is okay to close the case and Nutanix Support will close the case.

Document ID:HT516497
Original Publish Date:05/17/2024
Last Modified Date:05/23/2024