You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I made a mistake. There is no big problem with the allocation logic of tcp_receiver. However, the load calculation uses the pt_get_1m_proc_load function, which is not accurate enough and can easily cause CPU spikes. This problem can be solved by changing to pt_get_pt_proc_load
Running environment.
version: opensips 3.4.7 (x86_64/linux)
flags: STATS: On, DISABLE_NAGLE, USE_MCAST, SHM_MMAP, PKG_MALLOC, Q_MALLOC, F_MALLOC, HP_MALLOC, DBG_MALLOC, FAST_LOCK-ADAPTIVE_WAIT
ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
poll method support: poll, epoll, sigio_rt, select.
git revision: 79ebc9d
main.c compiled on with gcc 10
Describe the problem
After upgrading from 3.2 to 3.4, the tcp receiver distribution becomes unbalanced. proc_load will have spike

3.2 version
3.4 version

Describe alternatives you've considered
There may be a problem with the following commit
5a6399b
The text was updated successfully, but these errors were encountered: