IBM Support

Must Gathers when Engaging Support in PureData System for Analytics

Question & Answer


Before I raise a PMR, what logs should I capture to assist support in resolving my problem as quickly as possible?


Must Gathers when engaging Support:

For IDAA customers, please consult your documentation for tracing options that mimick below behavior:


1. log into active node as nz, check nzhealthcheck version

$nz /nz/kit/bin/adm/nzhealthcheck -v

a. If version is below 2.3, download latest standalone version:

Download and install latest Software Support Tools from fix central
(they are backwards compatible within 7.X releases)

Run in standalone mode:

$nz /nz/support/contrib/bin/adm/nzhealthcheck/nzhealthcheck -S

Enter root password when prompted

b. If version is at or above 2.3:

In both cases, the result is a .txt file which should be attached to the PMR


1. Log into active node as root, check nzlogcollector version

$root /nz/support/bin/adm/nzlogcollector/nzlogcollector -v

If version is below

Download and install latest Software Support Tools from fix central
(they are backwards compatible within 7.X releases)

If you are unsure what logs to capture, run nzlogcollector without any flags for a cursory exam:

$root /nz/support/bin/adm/nzlogcollector/nzlogcollector -outputdir /nzscratch/

The command will create a .zip file. Please attach this zip file to the PMR.

Under the direction of support, you may invoke a suite of target area component based collection:

nzlogcollector flags (use outputdir flag for all commands and be mindful of space considerations)

nzlogcollector flags
Collects logs to narrow down host hardware issues
-component host rack1.host1

-component host rack1.host2
Collects logs that impact cluster failover and high availability
-component cluster
Collect logs across an entire H chasis
-component spa spa1

-component spa spa2
Collects logs targeted to a particular SPU
-component spu spa1.spu1

-component spu spa1.spu3
Collects logs pertaining to both internal and client facing networking
-component network
Collects logs pertaining to disk access and sas fabric health issues
-component sasphy spa1

-component sasphy spa2
Collects logs pertaining to specific disk
-component disk spa1.diskEncl1.disk1
Collects logs which impact all tables across a specific database
-component database testDB <DBname>
Collects logs for a specific query based on plan file ID
-component plan 1234 <PLANID>

Similarly, the following are “symptom based” flags which may be issued:

SymptomWhen to runnzlogcollector flags
When the database has crashed/restarted.
Note: cores may be collected, be conscious of space concerns
-symptom crash
queries are not progressing; many basic commands will hang or time-out-symptom npshang
Queries are not progressing; query are waiting on a specific dataslice-symptom spuhang

[{"Product":{"code":"SSULQD","label":"PureData System for Analytics"},"Business Unit":{"code":"BU001","label":"Analytics Private Cloud"},"Component":"Not Applicable","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"1.0.0","Edition":""}]

Document Information

Modified date:
16 June 2018