Is placement Groups mandatory to setup Teradata developer free tier on AWS

Teradata Database on AWS
Enthusiast

Is placement Groups mandatory to setup Teradata developer free tier on AWS

Hi All,

While trying to deploy Teradata developer free tier instance on AWS, it is asking for "placement groups", 

Is "placement groups" is mandatory to setup a developer system with 2 nodes?

 

6 REPLIES
Teradata Employee

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

For any multi-node Teradata instance (even 2 node developer), you need high speed / low latency communications between the nodes.

That's the purpose of the cluster placement group.

Enthusiast

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

Thank you Fred.

Enthusiast

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

Hi Fred,

 

We successfully deployed teradata instance on AWS. But after some time, restart generated by Bynet driver. We tried to bring the system up with restart  but unsuccessful.  Could you please help/provide steps to bring the system online. 

 

Below is the snippet from /var/log/messages,

 

Jul 6 03:21:28 SMP001-01 kernel: klogd 1.4.1, ---------- state change ----------
Jul 6 03:21:28 SMP001-01 Teradata[26214]: INFO: Teradata: 13014 #Event number 33-13014-00 (severity 20, category 3) TPA reset generated by Bynet driver.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: CPU[ALL]: Throttling NOT Supported! Failed To Set to 1000.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: Metering Disabled.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: CPU[ALL]: Throttling NOT Supported! Failed To Set to 1000.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: Metering Enabled.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: meter_mode Set To 0.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: CPU[ALL]: Throttling NOT Supported! Failed To Set to 1000.
Jul 6 03:21:28 SMP001-01 kernel: tdmeter: Metering Disabled.
Jul 6 03:21:28 SMP001-01 kernel: INFO: BLLI: 140002603 #online with 1 nodes.
Jul 6 03:21:43 SMP001-01 Teradata[26214]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.
Jul 6 03:21:47 SMP001-01 kernel: INFO pdesys: sch_kiosched_exit: PSF I/O Scheduler exited


#### egrep "lost contact|online with" /var/log/messages

 

Jul 6 01:52:26 ip-172-31-0-130 kernel: INFO: BLLI: 140002603 #online with 2 nodes.
Jul 6 03:21:28 SMP001-01 kernel: WARNING: BLLI: 140002903 #lost contact with node 1-2, restart initiated.
Jul 6 03:21:28 SMP001-01 kernel: INFO: BLLI: 140002603 #online with 1 nodes.
Jul 6 03:23:26 SMP001-01 kernel: DEGRADED: BLLI: 140003703 #lost contact with all nodes on eth0:0-udp-1001-172.31.1.140.
Jul 6 03:46:27 SMP001-01 kernel: INFO: BLLI: 140002603 #online with 1 nodes.

Teradata Employee

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

Hello Shankar,

Normally, in this case, Node Failure Recovery (NFR) should occur and the faulty node should get replace with new node. Could you please check how many instances are running in EC2?

Also if possible, could you please provide the below logs?

 

1.)  /var/log/messages

2.)  /var/log/tdc*   // Logs starting with "tdc"

3.)  What does the "pdestate -a" command shows on both the nodes?

 

If possible please mail the logs to abhijeet.shah@teradata.com

 

Thanks,

Abhijeet K Shah

 

Enthusiast

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

Thank you Abhijeet.

 

We are  running Teradata on EC2 instance with two nodes  and both nodes  are on "PDE state: DOWN/HARDSTOP". We sent logs to you via e-mail. Appreciate if you can check and provide us the solution/steps to bring the system back online. Bye the way we are using Teradata developer free version on EC2 instance.

 

Thanks in advance.

 

Teradata Employee

Re: Is placement Groups mandatory to setup Teradata developer free tier on AWS

// In your case as per the logs, no new node spun-up when the node 001-02 lost the contact and split net condition detected.
 - One of the reason for Node failure recovery(NFR) was not kicked-in for node failure/lost contact is might related to instance limits, please check.
 - Teradata recommends increasing the limits to one more than the size of the system to support Node Failure Recovery.

 

// You might aware of the split net condition (KCS008531) and even with 90 secs bynet timeout, failed to establish communication with node 001-02 after multiple attempts of the system reboot. Suggesting to check with AWS Support to re-establish the network on node 001-02 and re-deploy the system, if you are unable establish the network in any case.
 
Jul  6 03:21:28 SMP001-01 kernel: WARNING: BLLI: 140002903 #lost contact with node 1-2, restart initiated.
Jul  6 03:21:28 SMP001-01 kernel: klogd 1.4.1, ---------- state change ----------
Jul  6 03:21:28 SMP001-01 Teradata[26214]: INFO: Teradata: 13014 #Event number 33-13014-00 (severity 20, category 3) TPA reset generated by Bynet driver.
Jul  6 03:21:28 SMP001-01 kernel: INFO: BLLI: 140002603 #online with 1 nodes.
Jul  6 03:21:43 SMP001-01 Teradata[26214]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.


Jul  6 03:38:07 SMP001-01 pcl[9053]: DEGRADED: TdatTools: 29003 #Node which failed to connect: "byn001-02"
Jul  6 03:38:07 SMP001-01 pcl[9053]: WARNING: TdatTools: 29002 #1 Node of 2 failed to connect
Jul  6 03:38:08 SMP001-01 Teradata[9074]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.


Jul  6 03:43:12 SMP001-01 shutdown[10020]: shutting down for system reboot
Jul  6 03:48:08 SMP001-01 pcl[7308]: DEGRADED: TdatTools: 29003 #Node which failed to connect: "byn001-02"
Jul  6 03:48:08 SMP001-01 pcl[7308]: WARNING: TdatTools: 29002 #1 Node of 2 failed to connect
Jul  6 03:48:38 SMP001-01 recond[7294]: WARNING: TdatTools: 29002 #BOOT-TIME: conditionally waiting up to: 30 more seconds for node connections
Jul  6 03:49:11 SMP001-01 Teradata[7402]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.


Jul  6 06:45:47 SMP001-01 kernel: INFO: BLLI: 140001003 #blmtune changed blli_sar_down_wait from 10000000 to 90000000.
Jul  6 06:48:32 SMP001-01 pcl[19449]: DEGRADED: TdatTools: 29003 #Node which failed to connect: "byn001-02"
Jul  6 06:48:32 SMP001-01 pcl[19449]: WARNING: TdatTools: 29002 #1 Node of 2 failed to connect
Jul  6 06:48:34 SMP001-01 Teradata[19470]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.
Jul  6 06:56:33 SMP001-01 pcl[21558]: DEGRADED: TdatTools: 29003 #Node which failed to connect: "byn001-02"
Jul  6 06:56:33 SMP001-01 pcl[21558]: WARNING: TdatTools: 29002 #1 Node of 2 failed to connect
Jul  6 06:56:34 SMP001-01 Teradata[21578]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.


Jul  6 07:00:45 SMP001-01 shutdown[22646]: shutting down for system reboot
Jul  6 07:02:10 SMP001-01 kernel: INFO: BLLI: 140001003 #blmtune changed blli_sar_down_wait from 10000000 to 90000000.
Jul  6 07:03:52 SMP001-01 pcl[4299]: DEGRADED: TdatTools: 29003 #Node which failed to connect: "byn001-02"
Jul  6 07:03:52 SMP001-01 pcl[4299]: WARNING: TdatTools: 29002 #1 Node of 2 failed to connect
Jul  6 07:04:22 SMP001-01 recond[4285]: WARNING: TdatTools: 29002 #BOOT-TIME: conditionally waiting up to: 30 more seconds for node connections
Jul  6 07:04:55 SMP001-01 Teradata[4392]: INFO: Teradata: 13006 #Event number 33-13006-00 (severity 10, category 12) split net detected.
Jul  6 07:07:07 SMP001-01 kernel: WARNING: BLLI: 140003603 #cannot contact any nodes on eth0:0-udp-1001-172.31.1.140.