Vertica - Networksend Issue while using COPY FROM / EXPORT

suryateja

Hi Team,

We are facing an issue with the data movement from one node vertica db to 3 node vertica db in aws. We are able to connect the db using CONNECT TO and after running the EXPORT TO it is getting failed as below.

ERROR: [vertica] NetworkSend on v_vertica_node0001: failed to open connection to node v_vertica_node0001 (socket error: Connection timed out)

When we tried to telnet the server from each other they are getting connected and can someone help on this .

Regards,
Suryateja.

VValdar

Hi Suryateja,

Maybe you should try COPY FROM VERTICA:
https://docs.vertica.com/12.0.x/en/data-export/db-export-and-import/copying-data-from-another-db/

suryateja

I tried that as well and seeing the same issue. Below is the log file

2023-04-13 16:29:49.517 Init Session:0x7f60f9ffb700-a000000000c5f7 [Txn] Begin Txn: a000000000c5f7 'COPY report.dim_device_info FROM VERTICA Vertica.report.dim_device_info'
2023-04-13 16:29:49.550 Init Session:0x7f60f9ffb700-a000000000c5f7 [Opt] In BulkLoad::loadFromVertica(): our encryption: 0, other encryption: 0
2023-04-13 16:29:49.551 Init Session:0x7f60f9ffb700-a000000000c5f7 [Opt] Fast loading constant values enabled
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [Opt] Dump SendRecvNodes Map:
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [Opt] Dump SendRecvNodes Map:
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [PROTOCOL] ClusterPlanner: node: 0, dbname : vertica, db: 0
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [PROTOCOL] ClusterPlanner: node: 0, dbname : vertica, db: 0
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [PROTOCOL] ClusterPlanner: Have a default address '10.201.3.235'
2023-04-13 16:29:49.552 Init Session:0x7f60f9ffb700-a000000000c5f7 [PROTOCOL] ClusterPlanner: Have a default address '10.201.3.235'
2023-04-13 16:29:49.561 Init Session:0x7f60f9ffb700-a000000000c5f7 [EE] DataTargetProxy allocates 32 receive buffers (buffer size = 1056768)
2023-04-13 16:29:49.561 Init Session:0x7f60f9ffb700-a000000000c5f7 [EE] DataTargetProxy allocates 32 receive buffers (buffer size = 1056768)
2023-04-13 16:31:58.870 Init Session:0x7f60f9ffb700-a000000000c5f7 @v_vertica_node0001: V1001/4054: [vertica] NetworkSend on v_vertica_node0001: failed to open connection to node v_vertica_node0001 (socket error: Connection timed out)
2023-04-13 16:31:58.873 Init Session:0x7f60f9ffb700-a000000000c5f7 [PROTOCOL] Client::rollupErrors: Num Errors: 1, Error Rollup: rootError 0, commError 1, ignoreCancel 1
2023-04-13 16:31:58.873 Init Session:0x7f60f9ffb700-a000000000c5f7 [VMPI] Cleaning network channels for plan a000000be243a7 (failed=1)
2023-04-13 16:31:58.902 Init Session:0x7f60f9ffb700-a000000000c5f7 @v_vertica_node0001: V1001/4054: [vertica] NetworkSend on v_vertica_node0001: failed to open connection to node v_vertica_node0001 (socket error: Connection timed out)

let me know what exactly is going wrong between these 2 servers

Bryan_H

For AWS connectivity, please check the security group rules between the clusters. My recollection is that you need other service ports open, notably TCP 5434 and TCP 5444, to support COPY FROM VERTICA. Please refer to the following list of ports and uses: https://docs.vertica.com/12.0.x/en/setup/set-up-on-premises/before-you-install/configure-network/ensure-ports-are-available/

SruthiA

are the clusters hosted on AWS in private VPC's? If so, make sure you have a export address which can help in accessing each node in other cluster when you run COPY from VERTICA