Fix timing issue in dual channel replication COB test

[enjoy-binbin]

After #2829, valgrind report a test failure, it seems that the time is not enough to generate a COB limit in valgrind.

[enjoy-binbin]

valgrind ci: https://github.com/valkey-io/valkey/actions/runs/19392859905

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 72.49%. Comparing base (86db609) to head (7bb3af9).
⚠️ Report is 2 commits behind head on unstable.

Additional details and impacted files

@@             Coverage Diff              @@
##           unstable    #2847      +/-   ##
============================================
+ Coverage     72.26%   72.49%   +0.22%     
============================================
  Files           128      128              
  Lines         70370    70371       +1     
============================================
+ Hits          50851    51013     +162     
+ Misses        19519    19358     -161     

see 14 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[zuiderkwast]

If 5 seconds is not enough in our valgrind builds, maybe it's not enough in some real scenarios too? Should we increase the default to 10 seconds?

[enjoy-binbin]

If 5 seconds is not enough in our valgrind builds, maybe it's not enough in some real scenarios too? Should we increase the default to 10 seconds?

This is actually a test dependency, in real wordl scenarios, servers that slow seem unlikely. It can't even generate 1100k of traffic in 5 seconds.
$primary config set client-output-buffer-limit "replica 1100k 0 0"

[zuiderkwast]

OK