cbcopy_helper write: broken pipe

[adnanhamdussalam]

Hi ,

I am facing below errors while using cbcopy. The firewall is disabled and IP Tables are also not configured.

20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640914-[DEBUG]:-Helper stopping
20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640914-[ERROR]:-cbcopy_helper exited with error: failed to redirect stream: readfrom tcp 192.168.3.71:52232->192.168.3.72:1072: write tcp 192.168.30
.17:52232->192.168.3.72:1072: write: broken pipe
20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[DEBUG]:-Connection to 192.168.3.72:1072 is closed.
20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[DEBUG]:-Helper stopping
20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[ERROR]:-cbcopy_helper exited with error: failed to redirect stream: readfrom tcp 192.168.3.71:52226->192.168.3.72:1072: write tcp 192.168.30
.17:52226->192.168.3.72:1072: write: broken pipe

Can you tell if the issue is related to cbcopy utility?

[adnanhamdussalam]

One more observation:

     Schema     |   Name    | Type  |        Owner         | Storage | Persistence | Access method | Size  | Description
----------------+-----------+-------+----------------------+---------+-------------+---------------+-------+-------------
 dev | my_calls | table | my_user | heap    | permanent   | heap          | 64 GB |
(1 row)

gpafiniti=# select count(*) from dev.my_calls;

count
-------
     0
(1 row)

source table is 64 GB and on target the size reached 64GB and also seeing the query is active in pg_stat_activity
and it is in active from last 4 hours
No errors in cbcopy logs or in db logs . The session get stuck.

Kindly need your support in the above issue.

[robertmu]

Hi ,

I am facing below errors while using cbcopy. The firewall is disabled and IP Tables are also not configured.

20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640914-[DEBUG]:-Helper stopping 20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640914-[ERROR]:-cbcopy_helper exited with error: failed to redirect stream: readfrom tcp 192.168.3.71:52232->192.168.3.72:1072: write tcp 192.168.30 .17:52232->192.168.3.72:1072: write: broken pipe 20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[DEBUG]:-Connection to 192.168.3.72:1072 is closed. 20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[DEBUG]:-Helper stopping 20251218:23:36:07 cbcopy_helper:gpadmin:VGX-GP:1640903-[ERROR]:-cbcopy_helper exited with error: failed to redirect stream: readfrom tcp 192.168.3.71:52226->192.168.3.72:1072: write tcp 192.168.30 .17:52226->192.168.3.72:1072: write: broken pipe

Can you tell if the issue is related to cbcopy utility?

Based on the helper's error logs you provided:

1. "broken pipe" means the destination-side helper has exited, causing the client-side helper to write data to a socket that the remote side has already closed. There could be many reasons why the destination helper exited — please refer to the README of the cbcopy and investigate why your destination-side helper exited.

2. "readfrom tcp 192.168.3.71:52232->192.168.3.72:1072: write tcp 192.168.30.17:52232->192.168.3.72:1072" indicates that your client-side helper has multiple source addresses when connecting to the destination helper. You may want to investigate why there are multiple source addresses from the client to the server, but this issue does not appear to be critical

[robertmu]

One more observation:

     Schema     |   Name    | Type  |        Owner         | Storage | Persistence | Access method | Size  | Description
----------------+-----------+-------+----------------------+---------+-------------+---------------+-------+-------------
 dev | my_calls | table | my_user | heap    | permanent   | heap          | 64 GB |
(1 row)

gpafiniti=# select count(*) from dev.my_calls;

count
-------
     0
(1 row)

source table is 64 GB and on target the size reached 64GB and also seeing the query is active in pg_stat_activity and it is in active from last 4 hours No errors in cbcopy logs or in db logs . The session get stuck.

Kindly need your support in the above issue.

This issue is unrelated to the previous "broken pipe" error. Based on your description, it only indicates that cbcopy and the database are both in a normal working state — the query is active and data has been transferred. Without more detailed information, we cannot diagnose further