[strongSwan] Issue with threads blockage in high load

[Chinmaya Dwibedy]

Hi,

I am doing scalability/load testing using 5.0.4 strongswan and load-tester plugin. I am debugging to figure out the potential bottleneck in high loads (25-30k IPsec connections). I find that, at IKE initiator ends, most of the threads are blocked forever in pthread_cond_timedwait (). Here goes the stack trace. Any idea what might be cause behind this? Please note that, upon profiling I did not find any costly function at both the ends. 

Thread 1 (process 1235):
#0  0x00000055630e4eb4 in pthread_cond_wait () from /lib64/libpthread.so.0
#1  0x0000005563141d44 in process_jobs (worker=0x126013900) at processing/proces
sor.c:278
#2  0x0000005563145034 in thread_main (this=<value optimized out>) at threading/
thread.c:309
#3  0x00000055630df698 in ?? () from /lib64/libpthread.so.0

Thread 1 (process 1244):
#0  0x00000055630e4eb4 in pthread_cond_wait () from /lib64/libpthread.so.0
#1  0x0000005563141d44 in process_jobs (worker=0x1260157b0) at processing/proces
sor.c:278
#2  0x0000005563145034 in thread_main (this=<value optimized out>) at threading/
thread.c:309
#3  0x00000055630df698 in ?? () from /lib64/libpthread.so.0

Regards,
Chinmaya