Diagnostics CLI Command

This page describes the diagnostics CLI command which is used for running cluster diagnostics.

Overview

The diagnostics CLI command is used for collecting and uploading diagnostic data about clusters and hosts for analysis by the Weka Support and R&D Teams in order to help with troubleshooting. There are two relevant commands:

  • weka diags used for cluster-wide diagnostics from any host in the cluster.

  • weka local diags used for running diagnostics on a specific host, which should be used if the host cannot connect to the cluster.

Collecting Diags

Command: weka diags collect

Use the following command to create diagnostics information about the Weka software and save it for further analysis by the Weka Support team:

weka diags collect [--id id] [--timeout timeout] [--output-dir output-dir] [--core-limit core-limit] [--clients] [--tar]

If the command is run with the local keyword, information is collected only from the host on which the command is executed. Otherwise, information is collected from the whole cluster.

Parameters in Command Line

Name

Type

Value

Limitations

Mandatory

Default

id

String

Specified ID for this dump

No

Auto-generated

timeout

String

How long to wait when downloading diags from all hosts

format: 3s, 2h, 4m, 1d, 1d5h, 1w

0 is infinite

No

10 minutes

output-dir

String

Directory to save the dump to

No

/opt/weka/diags

core-limit

Number

Limit to processing this number of core dumps, if found

No

1

clients

Boolean

Collect from client hosts as well

No

No

tar

Boolean

Create a TAR of all collected diags

No

No

Note: In the following situations the local option should be used: no functioning management process in the originating host or the hosts being addressed, no connectivity between the management process and the cluster leader, the cluster has no leader, the local container is down, the host cannot reach the leader or a remote host fails to respond to the weka diags remote command.

Uploading Diags to Weka Home

Command: weka diags upload

Use the following command to create diagnostics information about the Weka software and save it for further analysis by the Weka Support team:

weka diags upload [--timeout timeout] [--core-limit core-limit] [--dump-id dump-id] [--host-id host-id]... [--clients]

Parameters in Command Line

Name

Type

Value

Limitations

Mandatory

Default

timeout

String

How long to wait for diags to upload

format: 3s, 2h, 4m, 1d, 1d5h, 1w

0 is infinite

No

10 minutes

core-limit

Number

Limit to processing this number of core dumps, if found

No

1

dump-id

String

ID of an existing dump to upload

This dump ID has to exist on this local machine

No

If an ID is not specified, a new dump is created

host-id

Comma-separated list of numbers

Host IDs to collect diags from

This flag causes --clients to be ignored

No

All hosts

clients

Boolean

Collect from client hosts as well

No

No

When running the command with the upload option, the information is uploaded to a Weka-owned S3 bucket, and an access identifier is provided as an output. This access identifier should be sent to the Weka Support Team, which will retrieve the information from the S3 bucket.

Note: The upload command requires Internet connectivity which allows HTTPS access to AWS S3 endpoints.

Note: The upload process is asynchronous. Therefore connectivity failures will be reflected in failure events on the event log, while the command will still exit successfully.

Weka Processes Traces

The Weka processes constantly create traces. These traces are rotating to consume only a defined capacity. When encountering issues, sometimes there is a need for Weka support to observe those traces and, in some situations, to prevent specific traces from rotating until resolution.

It is possible to control the traces retention settings using the various weka debug traces CLI commands.

Last updated