Incident Information Checklist | Teradata Vantage - Incident Information Checklist - Advanced SQL Engine - Teradata Database

If you notice your system is having problems and you want to contact Teradata Services to report an incident, collect as much information about your system as possible. The list below is not exhaustive and is meant only to help you begin gathering details.

Note that not all of the questions may be applicable to your system and that other aspects of performance problems may not be addressed below. It is important to provide any other information that is not listed below but that you feel is important.

1. Describe any performance problems and its impact.
   • Is the system hung?
   • Is there a general system-wide slowdown?
   • Is there a slow query or class of queries?
   • Has there been a gradual deterioration?
   • Are there any time-outs?

     Example: The entire system runs fine for a while, then performance gradually deteriorates over a span of a few days. Marketing XYZ queries time out more often until eventually the entire system hangs. A Teradata reset clears the problem.

2. Is performance degraded for the entire system or part of the system?
   • A particular node(s) or clique(s)?
   • A particular subsystem like BYNET or disk array?
   • A particular application or client?
   • The network?
   • A particular group or class of users?
   • A particular job or query?
   • Do you have a performance baseline to compare against?

     Example: The Marketing XYZ queries, when run from a gateway connected session via a Web interface, had degraded response time from 10 minutes to approximately 30 minutes.

3. When does the problem occur?
   • All the time?
   • A particular time of day, day of week, week of month, month of year?
   • When did this problem start? Did it start suddenly?
   • Is it getting progressively worse?
   • Is the performance problem intermittent and unpredictable?
   • Has this problem happened before?
   • Are there any other past incidents that might be relevant to the current performance situation at this site?
   • Is the problem reproducible? If so, how?
   • What else is running while this problem occurs?
4. Other than performance, do you observe any other symptoms? For example:
   • Any unusual errors, warnings or messages, especially lately?
   • MiniCylPacks?
   • File space problems?
   • Memory depletion?
   • Lock contention?
   • Network congestion, collisions, bottlenecks?
   • z/OS errors, cell problems, in particular?
   • Are users having problems logging into the system and the database?
   • Are any third party incidents currently open?
5. What recent changes have been made to the system?
   • Recent configuration changes?
   • New hardware or new software?
   • New applications or new queries?
   • Different job mix and scheduling?
   • Have you just newly switched to a production environment?
   • Increased or different workload?
   • New or additional users?
   • New or additional data?
   • System tuning changes?
   • Changes to workload management methods?
6. Describe the data processing environment and where Teradata fits into the context.
   • What is your system generally used for? (Decision Support, OLTP, OLAP, Ad-Hoc Queries, a mix?)
   • What other vendor applications are you using?
   • Which applications are your most critical applications?
   • Who are the users? Are there different groups of users running different applications?
   • What types of data and how often is data loaded?
   • Can you provide a high-level block diagram illustrating the Teradata system, interfaces, networks, hosts, data marts, intermediate Servers, clients, sessions, users, ODBC, and so on?
7. Describe the Teradata System Configuration.
   • What are the versions of DBS, PDE, TCHN, TGTW, TDGSS, BYNET, and storage devices?
   • What is the number of nodes, CPU speed, AMPs per node, memory size per node, and storage per AMP?
   • Have the workload management methods been set up differently than the default settings?
   • Are there any unusual or asymmetrical aspects to the system configuration, such as different node types, different CPU speeds, different size cliques, different numbers of VPROCs per node, different storage subsystem, BYNET or other I/O components?
8. What supporting evidence and data do you have to assist Teradata Services in troubleshooting?
   • Do you have logs for: resource usage data, access logging, the Viewpoint Lock Viewer portlet, system accounting, syschk and mboxchk, blmstats, tdnstats, event logs and other error logs.
   • Do you have baseline data?
   • If you are trying to troubleshoot a particular query, have you measured performance of the same query from different clients (that is, BTEQ versus network attached client or have you tried to isolate bottlenecks)?
   • Has there been network congestion or collisions?