Recent comments
Teradata Manager - Where is that in Viewpoint?
The Teradata Viewpoint team is often asked how do I do this Teradata Manager function in Viewpoint? Although Teradata Viewpoint is fairly intuitive and offers excellent on-line help, it seemed worth creating an article related to this subject to accomplish two things. First, provide some high level guidance of what portlets may want to be considered for general Teradata Manager functionality. Secondly, provide a discussion forum for more in-depth Teradata Manager feature questions.
Note there is general agreement that a full Teradata Manager feature by feature translation guide would become too large and cumbersome to be effective. So keep reading if this is of interest to you and contribute accordingly with any items you would like discussed.
This article will go through the high level breakdown of Teradata Manager high level feature groups with indications of what Viewpoint portlets should be considered for that functional equivalence. Pertinent to this discussion are the existing articles on the Teradata Viewpoint 13.03 release and subsequent Teradata Viewpoint 13.03.00.04 patch release, both of which provide important information on the Teradata Manager transition to Viewpoint and Teradata Management Portlets. The Teradata Viewpoint 13.02 and Teradata Viewpoint 13.10 release articles provide information on the TASM / workload management portlets. All of these articles should be read if you have not done so already. Lastly, make sure you review the latest Teradata Viewpoint Configuration Guide as this has additional Teradata Manager transition information.
The Teradata Manager feature functionality tools are presented in four main feature groups:
As seen below in the Teradata Manager toolbar. Other items shown are for general navigation, formatting, etc.
Before diving into each of the feature groups, note that "Help" is replaced in Teradata Viewpoint with the built in on-line help implementation. Help in Viewpoint is found in the upper right corner and identified by the quetion mark icon shown below.
Now that help is covered, let's go through each of the feature function groups.
The "Monitor" menu is comprised of the following choices:
The "Dashboard" is documented as displaying the following information. Each item is shown with the associated Viewpoint portlet or portlets in italics that provide similar information. Overall, the health of a Teradata system is defined in Viewpoint through the System Health portlet.
- Virtual utilization and Physical utilization: Node Resources
- Sessions: System Health or Query Monitor
- Transaction rate: Query Monitor or Todays Statistics or Capacity Heatmap or Metrics Graph
- Response time: Todays Statistics
- I/O per transaction: Capacity Heatmap or Metrics Graph
- Recent history graphs showing trends: Capacity Heatmap or Metrics Graph
- Workload snapshots and history: Workload Health or Workload Monitor
The "Sessions" and "Session Information" provided similar functionality, all of which is now offered in the Query Monitor portlet. As a feature addition, personal session information for any particular user can be displayed in the My Queries portlet.
The "Nodes" and "Vprocs" information primarily is replicated in the Node Resources portlet although trend information for some of this information is also available in the Capacity Heatmap or Metrics Graph trending portlets.
The "TDWM" information was specific to TASM and workload management monitoring and reporting. All of this information is now available in the Workload Health or Workload Monitor portlets.
Last but certainly not least of the monitoring tools is "Performance Monitor" or maybe better known as just "PMON". PMON is described as providing performance data on four sets of data listed below with the associated Viewpoint portlet or portlets replacements listed in italics.
- General health information (configuration and summary): System Health
- Resource information (nodes and vprocs): Node Resources
- Session Information (statistics and lock information): Todays Statistics and Lock Viewer
- Historical session information: Query Monitor
Note that the monitoring rates for data collection some of which are specified in PMON are covered in the Viewpoint Admin menus under the Teradata Systems choice. Review the Data Collectors and Monitor Rates options.
The "Investigate" menu is composed of the following choices. Note the included migration status of each:
- Audit Log: Not migrated due to lack of usage and other log options.
- Error Log: Not migrated as this was only available for MP-RAS OS systems.
- Space Usage: Replaced by Space Usage portlet.
- Logon/Logoff History: Query the dbc.logonoff table or utilize the Query Monitor portlet.
- Locking Logger: Replaced by Lock Viewer portlet.
- Alert Viewer: Replaced by Alert Viewer portlet.
- System Emulation Tool (TSET): Remains a TTU product offering accessible from the Windows Start menu.
- Visual Explain: Remains a TTU product offering accessible from the Windows Start menu.
The "Analyze" menu is comprised of the following choices. Note the included migration status of each:
- Trends: See below.
- Resource History: Capacity Heatmap, Metrics Graph, and Metrics Analysis portlets.
- Statistics Wizard: Remains a TTU product offering accessible from the Windows Start menu.
- Index Wizard: Remains a TTU product offering accessible from the Windows Start menu.
The "Trends" menu choice provides resource usage trending which like the "Resource History" menu can now be found in the Capacity Heatmap, Metrics Graph, and Metrics Analysis portlets. The "Trends" menu also provides trend information for workload definitions that can be found in the Workload Monitor and Workload Health portlets. Regarding DBQL reporting, there will continue to be a need for the Database Query Log. However the Viewpoint DCS is a data repository with similar data, just not to the extent of DBQL, although it provides a vast majority of the value in a much easier access interface.
The "Administer" menu also has a variety of choices to consider. These are included in the following list where some have the migration status included on the line with others requiring more explanation and included below the list.
- Teradata Manager: See below.
- Teradata Database: Link to Teradata Administrator, a TTU product accessible from the Windows Start menu.
- Database Console: Replaced by Remote Console portlet.
- Alert Policies: Found in the Viewpoint Admin menu under Teradata Systems and the Alerts option.
- SQL Assistant: Remains a TTU product but is also complimented by the SQL Scratchpad portlet.
- BTEQ: Remains a TTU product offering accessible from the Windows Start menu.
- Teradata Database Setup: Similar setup for Viewpoint is available under the Teradata Systems menu.
- Statistics Collection: Functionality available through the Statistics Wizard TTU product.
- Move Space: Use the Space Usage portlet or the Teradata Administrator TTU product.
- Move Database: Functionality available through the Teradata Administrator TTU product.
- Workload Management: See below.
The "Teradata Manager" option is used to configure Teradata Manager, set up alerts, and data collection. This is done for Viewpoint in the Viewpoint Admin menus and in particular under the Teradata Systems and Alert Setup menu choices.
The "Workload Management" choice offers four options related to workload management.
- Dynamic Workload Manager: Replaced by Workload Designer portlet.
- Workload Analyzer: Remains a TTU product offering.
- Priority Scheduler (PSA): See below.
- Query Scheduler: Remains a TTU product offering.
Regarding the Priority Scheduler Administrator or often referred to as just PSA, a business decision was made to not bring that forward into Teradata Viewpoint. However all priority scheduler configuration operations can still be done via the "schmon" command. To make transition easier, one can use PSA to dump the set of schmon commands equivalent to each PSA profile one has.
So that is the high level overview of the status of Teradata Manager transition to Viewpoint. As mentioned in the introduction of this article, this is not a one by one feature transition but provides direction on where to start looking. This also allows for additional detailed discussions, as necessary, through comments to this article.
Hope this is helpful.
Hello Gary,
Thanks for this "Teradata Manager" and "ViewPoint" comparison article.
At present in Teradata Manager, if we go to "Remote Console" then we can redirect output to file.
Remote Console --> Options --> Redirect Output...
This is helpful is many cases. I don't see this functionality in "Remote Console" portlet in ViewPoint. Can you please tell me if this is cancelled or Am I missing anything while checking remote console portlet?
I can see there are many differences in look for "Workload Designer portlet" and "Dynamic Workload Manager". Can you please document high-level points for comparison between "Dynamic Workload Manager" and "Workload Designer portlet"?
Thanks & Regards,
Shrinivas Sagare
Thanks & Regards,
Shrinivas Sagare
Java script based web engine.. write content to local file system is quite a challenge.. think about someone write a java script webpage and start dumping virus to your local pc.. how would you like that???LOL..
well, maybe need a trusted java applet embeded in pages or webstart app to redirect output.
For redirection, one can simply use copy and paste operations. Regarding Workload Designer, I would recommend some other data resources if you're interested in that portlet and how it compares. See below.
1) Review the Viewpoint 13.10 release notes on DevX: http://developer.teradata.com/viewpoint/articles/teradata-viewpoint-13-10-released
2) Review the documented details of Workload Designer within on-line help or the Viewpoint User Guide PDF file version located under the Viewpoint Reference tab: http://developer.teradata.com/viewpoint/reference
3) Consider taking the Introduction to TASM Portlets class. More information at:
http://developer.teradata.com/viewpoint/training/introduction-to-tasm-portlets
Manoj, See responses below.
Regarding #1: 5-10 second response times are not normal. These delays could be network issues, over populated pages, or something else. As this is not a general concern, I'd suggest getting with the GSC to evaluate the situation.
For #2: The Viewpoint Configuration Guide documents all the granting of permissions necessary for the Viewpoint user id. I'd recommend having Teradata CS assist in your configuration. The Config Guide can be found on DevX at: http://developer.teradata.com/viewpoint/reference
The Alert Request table is a temporary queue so this is working as designed. Once Viewpoint processes the items in the queue with storage into Viewpoint, it purges the table entries.
For #3: On-line help search capability has been on the new features list for sometime but has been deferred as a lower priority as there is an alternate method for doing so. One can search on the PDF version if need be, this is the Viewpoint User Guide. The latest of which can be found on DevX at the same location as the Config Guide above.
Thanks Gary. I saw your reseponse just now. I moved the topic to the forum as i realised that i posted my questions in a wrong place for discussion.
Again thanks. I got the Config guide. let me retrieve the info. from there.
Manoj,
Can you re-publish your posting in this thread? I thought your comments were fine for this forum topic and now my response is confusing as what I was responding to is no longer listed here.
Gary
Gary,
I have published the deleted post.
************************************************************************************************
We are implementing Viewpoint and so far everything is good with its usage.
Being a newbie i have some basic questions.
1. I feel Viewpoint is slow considering the web application in a LAN Environment. I do understand the delay while fetching data. But even the settings (admin,profile,etc.) page and creation of new page is slow. Is this an isolated incident at my end or is this how the application behaves. The response time varies around 5-10 secs.
2. The Viewpoint id in Teradata has only read access to many of the admin related databases. Is there a reason for this. As a trial i tried giving full access to one of the table called "Alert Request" in the "DBCMNGR" database. To my surprise it purged out all the data from the table within 30 minutes. I tried investigating but could not able to find a reason. Any comments on why does it behaved so.
3. The Online help of Viewpoint answered most of my questions. But it will be more helpful if an option to search the terms is made available. This is a suggestion.
************************************************************************************************
Regards,
Manoj.
What would be the equivalent for the priority tab in TM for an appliance, in a VP portal.
Could you elaborate a bit more? What do you mean when you mention the priority tab in Teradata Manager?
In Teradata Manager --> Sessions options, there is a tab for showing the priority of a session. This is for the system which does not have TASM(Full scale, like an Appliance 25xx) and just the old Priority Scheduler configuration. Within Viewpoint, there is only the workload field/column available to view the category of a session's current status.
We setup TWDM workload throttle and object throttle, then we use TM monitor "Delay queue statistics" and "Object Query Statistics" to check count of the active, Limit and Delay, from viewpoint, where portlet can we see the throttle "Limit" settings.
To see the priority of a session, you can use Query Monitor and the "By Account String" option that should show the originating account string and $ priority indication.
Regarding TDWM throttles, one can use the Workload Monitor portlet to view overall system throttle information, and the Workload Health portlet for statistics on specific workload definitions (WDs). The Workload Designer portlet that replaces TDWM for all configuration operations can be used to view all workload management settings, including throttles. To learn more about these portlets, I'd suggest considering the Introduction to TASM Portlets training session:
http://developer.teradata.com/viewpoint/training/introduction-to-tasm-portlets
Gary,
In Teradata Manager under Alert Policies there is a tab for Events, Where can I set the action for Events in viewpoint.
Thanks,
Naveen
Actions are a three step process. The first step is to define the alerts infrastructure. For instance setting up SMTP for email or SNMP. This is done in the Admin -> Alert Setup menu.
The second step to set up events that trigger alerts is done in the Admin -> Teradata Systems -> Alerts. For example having an email sent to the DBA if System Health reports a critical condition for more than two minutes.
The third part of alerting is the Alert Viewer portlet for reviewing triggered alerts. For more information on the alerting process, check out the Viewpoint 13.03 release notes at the following link:
http://developer.teradata.com/viewpoint/articles/teradata-viewpoint-13-03-released
Worth mentioning, check out the Viewpoint Configuration Guide also as it has a full chapter on configuring alerts.
Viewpoint Alerts:
Mail Alerts are without the object name in the subject line,like the database name for a database space alert and user name for a session alert.
Example: A generic database space alert which is set sends the alert without the databasename on the subject line.
The message on the subject line is:
[Alert] EDWPROD - (Source: Viewpoint, Type: databaseSpace)
where:
- Alert Name
databasespace - Alert Type
Tried the Config. Guide but could not get a hint.
Is there a way to configure the alert to send the object name in the subject line while sending mail?
Customization of alerts is a roadmap consideration that has both Viewpoint and Database implications.
So the alerts of viewpoint at present does not show the object names in the subject line of the mail?
I think I misunderstood what you were initially asking. I just tested a few Viewpoint alerts with an email action and the resulting email lists the Teradata DB name in the subject line. If you find particular alerts that don't, then you may want to submit an enhancement request though the GSC.
LDAP in Viewpoint:
I configured LDAP for Viewpoint and it works fine for me. How ever, while logging through viewpoint if a wrong password is given once, the account gets locked on the first attempt.
I checked the log of the LDAP server and realised that viewpoint is making repeated attempts to logon (though it is an one time attempt from the user end) . i have shared a piece of the log below. First time i logged using the right credentials and logged off, the second time i gave a wrong password on purpose and this the time after 5 attempts the account is getting locked.
Note: Our Company Policy allows 5 attempts before the account gets locked out.
-----------------------------------------------------------------------------------------------------------------------------------------
Receipt_Time||Message||Name
-----------------------------------------------------------------------------------------------------------------------------------------
Jan 07 2011 10:06:50||User is currently locked out.||Logon attempt
Jan 07 2011 10:00:50|| ||User Account Locked Out
Jan 07 2011 10:00:51||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:50||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:57||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:51||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:57||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:51||User name is correct but the password is wrong.||Logon attempt
Jan 07 2011 10:00:20||Network: A user or computer logged on to this computer from the network.||User Logoff
Jan 07 2011 10:00:20||Network: A user or computer logged on to this computer from the network.||Successful Network Logon
Jan 07 2011 10:00:20|| ||Logon attempt using explicit credentials
-----------------------------------------------------------------------------------------------------------------------------------------
Is there anyone who has implemented LDAP in Viewpoint and faced this kind of an issue?
We used to have Alerts for Database restart (event Code:034029000), Node down, Vproc down, Mini Cyl-packs (0340514100), how can we conifg the similar alert in viewpoint? thanks.
For Node and Vproc , you can set it up using the following path
Admin > Teradata Server (any server that you choose) > Alerts (under setup menu) > Node (Alert setup menu) .
Create a new alert and in the dropdown where you can see metrics, select 'Status' and define it as equal to 'D'.
This alerts when node status is down. The same has to be applied for Vproc too. (refer to viewpoint configuration guide 13.10 , page 81)
Yet to find a solution for other alerts which are event based like Disk & Memory,Mini-Cyl Packs and restarts
We use the Teradata Manager "WD Delay Queue List" and "Object Delay Queue List" to monitor queries delayed by workload throttles and object throttles. I've yet to find anything equivalent in Viewpoint. Viewpoint shows how many queues in total are delayed and that an individual query is delayed, but not why (WD or object throttle). Of course, we're still running 13.03. Hopefully equivalent portlets will be available in 13.10 and above.
Tom McGrath
Another thing I couldn't find is Alert at query level, so that I can config alerts like long running Skewed queries, PJI exceed throshold queries.
Hi Gary. Congrats on Viewpoint...a product that just keeps getting better. But as someone who uses Teradata Manager all day long every day, I have to say you can take my TD Mgr away when you pry my cold dead hands off the keyboard. There are two key TD Mgr features that I haven't seen yet in Viewpoint. And they are critical to me doing my DBA work.
#1, TD Mgr peeks live into the PMAPI buffer to get up-to-the-second query step status. When I'm looking at a live explain, every time I click Refresh I get the current step progress. That allows me to watch second-by-second progress so I can tell exactly what step a query dies in if it fails. On Viewpoint, Refresh just pulls the data from the DCS database. So you can see the explain step no more recently than the last DCS refresh interval.
#2, on TD Mgr or PMON I can abort entire sets of sessions very quickly. If there are 20 fastload sessions from a single user and I want to cancel them all it is really easy. Just highlight and click abort. In Viewpoint it takes forever because I have to do them one at a time and log in with my Teradata credentials.
Are there any plans to add these two critical capabilities to Viewpoint?
Sorry for delays in response. Let's tackle alert questions first. Specific alert generation is done in the Admin -> Teradata Systems -> Alerts selection. You will find options for System, Node, VProc, Database Space, Session, Canary Queries, and System Health. Alerts on query aspects (like Skew, etc) will be done under the Session option. System status like down through System Health option.
Regarding "WD Delay Queue List" and "Object Delay Queue List", have you reviewed the Workload Monitor and Workload Health portlets?
For live peeks into the PMAPI buffer, you will have to be patient as this should be addressed in Viewpoint 13.12 release in 2H 2011. As always, forward looking statements are subject to change.
For actions on multiple sessions, this was resolved with Viewpoint 13.03. At the summary screen of queries, you choose an action that can then be applied to multiple queries in one operation. So that's already there.
That's great news about adding the PMAPI peek Gary. I've been raising my concern about that at Partners and other meetings but never heard that it was being worked on. I'm definitely looking forward to seeing that feature.
I hadn't noticed the multiple session actions but we are on VP 13.10 already and I just tested it. Works great. That's a real time saver. Thanks to you and the dev team for all the hard work!
I finally did find the info on the WD and object delays in the 13.11 manual. But our Development system, running TD 13.10, is not running TASM/TDWM, while our Production system is running TASM but still on TD 12.0.3. And I found the footnote that the delay columns are only available with database 13.10. OK - RTFM to me.
Next question: Is there an Viewpoint-equivalent alert to the Teradata Manager "Restarts" event?
Tom McGrath
You can setup alerts against System Health conditions including "Down" or "Degraded" so that's how we would suggest doing that. This can be found under Admin -> Teradata Systems -> Alerts -> System Health
I'm trying to replicate the PMON charts in viewpoint but I'm having trouble producing the same measures. In PMON I display the minimum, average and maximum CPU at the node and amp level. The node CPU is the most useful as generally this is what limits the system, i.e. is any node is running at 100% busy then Teradata can't do any more work. The difference between the minimum and maximum CPU is an indication of wasted CPU due to skewing.
In viewpoint there are a number of different CPU measures but I can't get the values to be the same. The issue seems to be that we have 2 non-TPA nodes which have very low CPU usage. In PMON these nodes were ignored in the CPU calculations, but in viewpoint these nodes are included. We have 9 active TPA nodes, so the average CPU usgae in viewpoint shows as just over 9/11ths of the value in PMON, and the minimum CPU is always very low, as it is the CPU used on the non-TPA nodes.
I'm using the Node CPU Usage measure. Is this the correct CPU measure to use? If so is there any way of only including nodes with amps?
The Node CPU Usage is the average over all nodes. We can consider modifying that calculation to exclude non-TPA nodes.
Also note that the CPU Usage currently includes Wait I/O CPU time. This will likely be removed from the calculation in an upcoming release.
Both of those changes would be very useful. Viewpoint should show the same measures as Teradata manager does. We are trying to switch off TD manager, but it's not going to be possible if we can't replicate all functions in viewpoint. Currently if my manager asks me how busy teradata is I can't tell him from figures in viewpoint.
There are enhancement requests open for both of those issues, so they will be considered for future releases of Viewpoint. If you want to see the CPU usage per node, check out the Node Resources portlet. From that portlet you can get the CPU usage on each node.
Also, I took a second look at how the CPU is calculated for the system. Viewpoint only counts the CPU for nodes with an AMP count > 0. If that's not how it appears to be working on your system, please work with support to help troubleshoot the problem.
Is it optional or mandatory to logon using teradata credentials to abort query/session in viewpoint? The viewpoint user guide document(query monitor section) tells me it is optional login to abort query. But I don't find a way to turn it off. I thought since viewpoint provides permissions for abort query and change workload through within viewpoint roles set up, then we don't need to log on again using teradata credential.
The problem is we have enabled LDAP logon on our teradata database system. SO, how can I provide my users abort query function if teradata login is required in viewpoint? Anybody else has same situation?
In migrating the Teradata Manager 12 alerts to Viewpoint 13.11, the comments associated with the TDWM alerts in the Teradata Manager Alert Policy disappeared. These comments are the only user-intelligible information in the alert messages to tell users what the alert is about – nobody knows what “Exception=10” means, and the even the support staff would have to dig to find out.
I’ve looked through the Viewpoint 13.11 documentation, and I haven’t found anything about adding comments to the non-user-friendly TDWM alerts going through Viewpoint. The only possibility that I can find is to develop (and maintain) a customized alerting process to capture the TDWM alerts and repackage them in a more friendly manner. This seems like a regression in the way TDWM alerts are processed, compared to Teradata Manager. And the users who have seen the “raw” TDWM alerts via Viewpoint agree.
Am I missing something?
Tom McGrath
Regarding the question on Teradata credentials for query/session actions like aborting, providing logon credentials is mandatory. The User Guide and on-line help state that:
You must log on with a user ID that has permissions to abort, change priorities or workloads, or release queries. If you log out, close a portlet, or open a new portlet instance, you must log on again.
The optional part is once we have your credentials, we will not ask for them again in this portlet within this session. If you logout and log back in, we will ask again on the first operation within a portlet. Also if you authenticate in Query Monitor and then go do something in another portlet, we will request credentials again. In other words, logon is optional within a portlet, once you have already provided your credentials. So only optional after it is required :-)
As of March 25th, Viewpoint only supports TD2 logon. Teradata LDAP authentication will be released in the Viewpoint 13.11.00.04 patch that should be available next week or latest early April.
Regarding the question on TDWM alerts, the alert functionality was enhanced when the TD Manager alerting was migrated to Viewpoint. In TD Manager, the message was associated with the alert action. This meant that if you had an action such as "Email DBAs", the same message was included in every email regardless of whether the condition that was met was for sessions, system metrics, or TDWM alerts. In Viewpoint, the message is configured on the alert rule, allowing much greater flexibility in the messaging. As a side effect of this change, a custom message cannot be associated with TDWM alerts.
However, you should note that in Teradata 13.10, we added Workload Name, Exception Name and Exception Action to the message that is written to the AlertRequest table and ultimately delivered as the message body in the matching Viewpoint alert.
I schedule and run many DBA related bteq scripts on the TD MGR box. Is there currently a scheduling feature like this in ViewPoint? If not, do you have any recommendations on how to replace this functionality?
This was covered in the Viewpoint 13.03 release article:
http://developer.teradata.com/viewpoint/articles/teradata-viewpoint-13-03-released
Summary is we made a business decision to discontinue the Teradata Manager scheduler as every modern operating system offers time based scheduling. So the recommendation would be to use a Windows or Linux somewhere in your infrastructure to schedule jobs with "cron" or "at" mechanisms.
Hi Gary,
Do you know if the Alert Viewer portlet will have the capability to view Alerts longer then 1 week in one window (without using rewind)?
No plans to add another duration selection in Alert Viewer at this time. With rewind, you should be able to accommodate other time durations although it may take multiple operations.
We have 3 * 2650 System, But One node is down.
The Node Resources portlet in Viewpoint doesn't show the down status, I was found the remote console.
How am I supposed to be set?
Is this node a hot standby node?
We dont have HSN node, we have 3 2650 node
Well the Node Resources portlet should indicate if the node is down and in tests here, it works correctly. Please submit a GSC incident to investigate the root cause of why you are not seeing the correct behavior.
I want to see that awt information at Metrics Analysis, so I can see it. But I found that different information between resawt and awt info at viewpoint.
Comparison between a specific time, but the information is wrong.
example resawt show the 120 awt number at specific, so I had to rewind and checking the awt number. The number is quite small.
AWT information in Metrics Analysis is described as "Average number of AMP worker tasks in use on each AMP". I am not sure how that compares to the resawt information you mention. I'd suggest submitting a question through a GSC incident for response.