]> git.ipfire.org Git - thirdparty/samba.git/commit
s4:torture/smb2: add 'smb2.bench.session-setup'
authorStefan Metzmacher <metze@samba.org>
Tue, 18 Jun 2024 14:44:37 +0000 (16:44 +0200)
committerVolker Lendecke <vl@samba.org>
Tue, 30 Jul 2024 15:18:42 +0000 (15:18 +0000)
commit7dc19dd94cb9c977ee61529543a841d1f7bc5259
treecb447756b6b667b8147488d79b6f23129eb28577
parent3bb6d441bf047bef6d95675057cecd3865a25540
s4:torture/smb2: add 'smb2.bench.session-setup'

This test loops over session-setup/logoff as fast as possible,
for nprocs * qdepth loops.

It was created to demonstrate the krb5 replay cache problem
(at least with MIT kerberos before 1.18, I've tested with
1.15 only).

This forces kerberos with the machine account:

time smbtorture //fileserver.example.com/ipc\$ \
        smb2.bench.session-setup \
        --option="torture:timelimit=600" \
        --option="torture:nprocs=50" \
        --option="torture:qdepth=1" \
        --machine-pass \
        --use-kerberos=required

This forces ntlmssp with the machine account:

time smbtorture //fileserver.example.com/ipc\$ \
        smb2.bench.session-setup \
        --option="torture:timelimit=600" \
        --option="torture:nprocs=50" \
        --option="torture:qdepth=1" \
        --machine-pass \
        --use-kerberos=off

This forces anonymous ntlmssp (without depending
on the DC performance):

time smbtorture //fileserver.example.com/ipc\$ \
        smb2.bench.session-setup \
        --option="torture:timelimit=600" \
        --option="torture:nprocs=50" \
        --option="torture:qdepth=1" \
        -U% \
        --use-kerberos=off

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue Jul 30 15:18:42 UTC 2024 on atb-devel-224
source4/torture/smb2/bench.c