Ricardo Nabinger Sanchez

Got it. On our end, we also face challenges when we have to adjust either of those (a lot more challenging for us when deadling with TE, though).

From the backtrace/code, it seems that this is the snippet of interest: ``` (gdb) p trash $5 = {size = 1048576, area = 0x7f3058211ed0 "strm=0x7f30585c8b90,145c0a src=x.x.x.x fe=ssl be=hyper_streaming dst=v01 txn=0x7f305858c750,43000...

Since `rate` was optimized out, here's `msg`: ``` (gdb) p msg $7 = 0x7f305857f710 "A bogus STREAM [0x7f30585c8b90] is spinning at 236832 calls per second and refuses to die, aborting...

I had gdb open/lingering just in case. I've enable pretty print to avoid having a horizontal scroll here in GH, so here we go: ```gdb (gdb) p *s $23 =...

Hey @capflam, no worries; here's the info you asked: ``` (gdb) p *((struct h1s *)s->scf->sedesc->se) $36 = { h1c = 0x7f3058526fe0, sd = 0x7f3058521150, flags = 1065104, subs = 0x7f30585c1490,...

Fix confirmed on `v3.0-dev11-53-gd33a5f8e1`. Thank you for your great work, as usual! :-)

This might have a slight pinch of the elusive #1895. The apparatus to detect is similar but our chosen handling was updated in case a core dump could give out...

While traversing the core dumps, we found one from the frontend (64 threads; 16 threads previously for the backend instance). It is much longer, so it has to be attached...

Hi Willy, We'll have to check 3.2, I think. As for the load, this server was around 40-50% of its peak load---which is odd, but also it's not at the...

Since there were multiple instances running, we get multiple core files. This one could be the "instance of interest", at least the backtraces are much more interesting. [core.haproxy.2045644.log](https://github.com/user-attachments/files/19764231/core.haproxy.2045644.log)