On Thu, Apr 4, 2024 at 2:36 AM Eelco Chaudron <echau...@redhat.com> wrote:
> > > On 3 Apr 2024, at 23:18, Ihar Hrachyshka wrote: > > > Before the patch, the size of the backlog depended on the type of socket > > (UNIX vs INET) as well as on the language (C vs Python), specifically: > > > > - python used backlog size = 10 for all sockets; > > - C used 64 for UNIX sockets but 10 for INET sockets. > > > > This consolidates the values across the board. It effectively bumps the > > number of simultaneous connections to python unixctl servers to 64. Also > > for INET C servers too. > > > > Signed-off-by: Ihar Hrachyshka <ihrac...@redhat.com> > > Hi Ihar, > > Thanks for submitting the patch. The patch looks fine to me, however, can > you elaborate on why you want to increase the size to 64? Is 10 giving > problems in specific scenarios? > > I guess I should've included it in the commit message, (and I am happy to send v2 with it updated), but... 1. Originally* the patch was implemented to allow more parallel fmt_pkt calls in OVN test suite (that rely on a python AF_UNIX unixctl server to transform scapy string format strings into byte strings). The problem with parallel handling of more than 10 unixctl AF_UNIX requests to python servers was noticed here: https://github.com/ovn-org/ovn/commit/0baca3e519756cbe98a32526ccc637bb73468743 2. Now, Brian also reports listen backlog issues in OpenStack environments for INET sockets, see: https://mail.openvswitch.org/pipermail/ovs-discuss/2024-April/053049.html * the patch was part of a series of patches - https://patchwork.ozlabs.org/project/openvswitch/list/?series=382739&state=%2A&archive=both - that generally improve AF_UNIX unixctl socket handling, which I plan to revive later, but nothing stops us from merging this before I get to it. > Cheers, > > Eelco > > Acked-by: Eelco Chaudron <echau...@redhat.com> > > _______________________________________________ dev mailing list d...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-dev
