I have verified this fix on noble-epoxy. I followed the testplan with one modification: Due to size limitations on my compute node, I created an amphora flavor with 24576 MiB of RAM.
When creating a pool on the listener thttps with ----------------------------- octavia-* 1:16.0.0-0ubuntu1~cloud1 ----------------------------- The pool's provisioning_status becomes ERROR. I observe haproxy triggering the oom-killer ----------------------------- [ 182.383148] haproxy invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 [ 182.383664] Out of memory: Killed process 6343 (haproxy) total-vm:13760176kB, anon-rss:4224kB, file-rss:3328kB, shmem-rss:10313344kB, UID:0 pgtables:20284kB oom_score_adj:0 ----------------------------- octavia-* 1:16.0.0-0ubuntu1~cloud2 ----------------------------- The pool's provisioning status is ACTIVE. No OOM occurs ----------------------------- ** Tags removed: verification-epoxy-needed ** Tags added: verification-epoxy-done -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2119987 Title: haproxy reload triggers OOM-killer for TERMINATED_HTTPS loadbalancers To manage notifications about this bug go to: https://bugs.launchpad.net/cloud-archive/+bug/2119987/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
