PostgreSQL pg_dump fails with "server closed the connection unexpectedly" during COPY operation

Problem Summary

My automated PostgreSQL backup script using pg_dump was working fine until a few days ago. Now it consistently fails with a connection error during the COPY operation on the alternate tables each time.

Environment Details

• PostgreSQL Version: 9.6.24
• Client: Running pg_dump via Docker container abc/postgres:9.6.24_0.3
• Database Server: fossology.com:1234
• OS: Linux (backup server) ubuntu dist
• Network: Corporate network with potential firewalls/proxies

Error Message

pg_dump: Dumping the contents of table "author" failed: PQgetCopyData() failed.
pg_dump: Error message from server: server closed the connection unexpectedly
    This probably means the server terminated abnormally
    before or while processing the request.
pg_dump: The command was: COPY public.author (author_pk, agent_fk, pfile_fk, content, hash, type, copy_startbyte, copy_endbyte, is_enabled) TO stdout;

Current Backup Script

#!/bin/bash
backup_path=/fossology_backup/pg_dumps
docker_image=abc/postgres:9.6.24_0.3
backup_file=fy_03.$(date +%Y%m%d_%H%M).gz
error_file=fy_03.$(date +%Y%m%d_%H%M).err

db_host=fossology.com
db_user=fly
db_name=fy
db_port=1234

docker run --rm -v $backup_path:$backup_path -v /root/.pgpass:/root/.pgpass $docker_image \
    pg_dump -h $db_host -U $db_user -d $db_name -p $db_port -F c \
    -f $backup_path/$backup_file 2> $backup_path/$error_file

What I've Tried

1. Connection testing: Basic psql connections work fine
2. Retry logic: Added retry mechanisms with delays - still fails consistently
3. Different pg_dump options: Tried various formats (-F p, -F d), verbose mode, different timeout settings
4. Database maintenance: Attempted VACUUM FULL

Additional Observations

• The backup worked perfectly until 3-4 days ago
• No recent changes to the backup script or environment
• The failure always occurs at talternate tables
• Was able to take the backup inside the docker container
• Backup is performed with NFS

Questions

1. What could cause PostgreSQL to suddenly start dropping connections during COPY operations?
2. Are there any PostgreSQL server-side settings or limits that might cause this behavior?
3. How can I diagnose if this is a table corruption issue vs. a network/timeout issue?
4. What alternative backup strategies would work around this specific issue?

Any insights into potential root causes or alternative approaches would be greatly appreciated!