Bug 1733709
Let's try on recent 7.7. We've made some changes recently about restarting. I will update this bug in a while. Created attachment 1594203 [details]
still present in RHEL7.7
nmcli conn add type ovs-bridge conn.interface brcnv
nmcli conn add type ovs-port conn.interface port0 master brcnv
nmcli conn add type ovs-port conn.interface brcnv-port master brcnv
nmcli conn add type ovs-interface conn.id brcnv-iface conn.interface brcnv master brcnv-port ipv4.method manual ipv4.address 192.168.100.154/24 ipv4.gateway 192.168.100.1 ipv4.dns 1.1.1.1
nmcli conn add type ovs-port conn.interface bond0 master brcnv ovs-port.bond-mode balance-slb
nmcli conn add type ethernet conn.interface eth1 master bond0
I see, the duplicate 'brcnv' interface name causes other issues on restart. I didn't realize before because I was using the script attached to the mail where the brcnv ovs interface is commented out. I'm debugging the problem. Working like a champ now. Huge thanks for the fix! Created attachment 1618946 [details]
Testing script for RHEL 7.6 / 7.7 / 7.8
TEST SETUP KVM based local VM Distro: RHEL 7.6 Component version: NetworkManager-1.12.0-6.el7 NetworkManager-ovs-1.12.0-6.el7 PREPARATION Add 1 Ethernet interfaces of type virtio the the testing VM. Set kernel parameters to disable Consistent Network Device Naming: # grubby --update-kernel=ALL --args="biosdevname=0 net.ifnames=0" # yum install -y NetworkManager-ovs # reboot TEST PROCEDURE 1. Run the testing script. 2. Get results. RESULT # nmcli con NAME UUID TYPE DEVIC eth1 6477ca5a-6986-43fe-9c09-6a361d7cf13a ethernet eth1 ovs-bridge-brcnv 3fd02a35-4c20-4ad8-bde5-5c6c7139f283 ovs-bridge brcnv ovs-slave-bond0 eae35fd6-62e9-487c-aae6-66c6613a9303 ovs-port bond0 ovs-slave-brcnv-port 49cd9a79-37ee-4a82-89a3-4e59d20aee4f ovs-port brcnv ovs-slave-port0 0ed8ffff-ecb9-41b4-b90b-a9dbb4a65d7b ovs-port port0 virbr0 d8e0628f-86cb-45fd-a89e-1a39fadc6018 bridge virbr brcnv-iface 20a45655-e480-4710-9990-7a110a63748c ovs-interface -- eth0 afe238f8-6895-486f-a898-1da642a2c120 ethernet -- ovs-slave-eth0 d4e8d943-4867-48b8-931e-09b14e8314e9 ethernet -- ovs-slave-eth1 d5c7e40f-82df-49a0-a833-44e7a212459c ethernet -- TEST SETUP 2 Distro: RHEL-7.8-20190912.3 Server x86_64 Component version: NetworkManager-1.18.0-5.el7_7.1 NetworkManager-ovs-1.18.0-5.el7_7.1 PREPARATION Make sure you have console connection to the testing machine. The SSH connection will be interrupted. TEST PROCEDURE 2 1. Create a Beaker job to reproduce the environment: https://beaker.engineering.redhat.com/jobs/3807638 2. Run the testing script. 3. Get results. RESULT 2 [root@qe-dell-ovs5-vm-59 bin]# ./bug_reproducer_vm.sh IP=10.16.134.65/23 GW=10.16.135.254 DNS=10.19.43.29 10.11.5.19 10.5.30.160 Connection 'ovs-bridge-brcnv' (d02b9bfe-ad58-46e8-ad29-cc32e0f5b8e1) successfully added. Connection 'ovs-slave-port0' (a607ddb1-54ee-4811-b163-63572739a220) successfully added. Connection 'ovs-slave-brcnv-port' (dc8f1916-6b4b-4401-a16d-0017b6904f10) successfully added. Connection 'brcnv-iface' (d5dd5975-c460-46c9-94a6-ae2290480fd9) successfully added. Connection 'ovs-slave-bond0' (99f105a3-bc66-4eae-b378-bbc2289099f4) successfully added. Connection 'ovs-slave-eth0' (4a7ba3f2-8b09-4378-a931-f3fd3ea33794) successfully added. Connection 'ovs-slave-eth1' (90d6d470-9aec-41de-97a8-d7317790e633) successfully added. Connection 'eth0' successfully deactivated (D-Bus active path: /org/freedesktop/NetworkManager/ActiveConnection/1) NAME UUID TYPE DEVIC ovs-bridge-brcnv d02b9bfe-ad58-46e8-ad29-cc32e0f5b8e1 ovs-bridge brcnv ovs-slave-bond0 99f105a3-bc66-4eae-b378-bbc2289099f4 ovs-port bond0 ovs-slave-brcnv-port dc8f1916-6b4b-4401-a16d-0017b6904f10 ovs-port brcnv ovs-slave-port0 a607ddb1-54ee-4811-b163-63572739a220 ovs-port port0 brcnv-iface d5dd5975-c460-46c9-94a6-ae2290480fd9 ovs-interface -- eth0 b42dda1e-b09e-4d30-a784-c2d3c1f64af9 ethernet -- eth1 78243c55-38db-4d81-82e8-c2fdc4babf47 ethernet -- eth10 e4a8b7fe-5beb-475d-9919-442e42c8af59 ethernet -- eth2 fde08446-2609-480c-9fe6-af8bf36c8ad7 ethernet -- eth3 db30f0ca-be62-40c2-ad12-2a3996388d25 ethernet -- eth4 ce3576b2-7a5a-46c2-8da9-94a5a8b1011b ethernet -- eth5 0d815d82-6b27-4e64-b005-8188c74b8d91 ethernet -- eth6 15a17c08-57b7-45a8-96eb-4eec8b944c45 ethernet -- eth7 1b59d383-1298-4005-b16f-c0b45d97f33c ethernet -- eth8 bddeab30-ce90-4659-ba8f-17de15a3a190 ethernet -- eth9 c43d6c36-ea66-467a-8e18-cf4083c95004 ethernet -- ovs-slave-eth0 4a7ba3f2-8b09-4378-a931-f3fd3ea33794 ethernet -- ovs-slave-eth1 90d6d470-9aec-41de-97a8-d7317790e633 ethernet -- NetworkManager-1.18.0-5.el7_7.1.x86_64 NetworkManager-ovs-1.18.0-5.el7_7.1.x86_64 Hi, the script works for me with NM 1.18.0-6 which is the last version available in RHEL 7.8. Can you please enable trace logging and attach NM logs? Thanks! Created attachment 1619342 [details] Output from the testing script TEST SETUP 3 Distro: RHEL-7.8-20190912.3 Server x86_64 Component version: NetworkManager-1.18.0-5.el7_7.1.x86_64 NetworkManager-ovs-1.18.0-5.el7_7.1.x86_64 TEST PROCEDURE 3 1. Create a Beaker job to reproduce the environment: https://beaker.engineering.redhat.com/jobs/3809424 2. Install Brew build NetworkManager-1.18.0-6.el7: https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=969104 3. To troubleshoot NetworkManager, follow the steps from: https://cgit.freedesktop.org/NetworkManager/NetworkManager/tree/contrib/fedora/rpm/NetworkManager.conf#n28 # sed -i "/RateLimitInterval/d" /etc/systemd/journald.conf # sed -i "/RateLimitBurst/d" /etc/systemd/journald.conf # echo -e "RateLimitInterval=0s\nRateLimitBurst=0" >> /etc/systemd/journald.conf # systemctl restart systemd-journald 4. Run the testing script. 5. Get messages from the system journal. RESULT 3 Expected result NOT achieved. Output is attached. Created attachment 1619343 [details]
System journal from reproducing the bug under RHEL-7.8-20190912.3
Created attachment 1619344 [details]
NM trace log from reproducing the bug under RHEL-7.8-20190912.3
Created attachment 1619522 [details]
Testing script for RHEL 7.6 / 7.7 / 7.8
Update the testing script.
Created attachment 1619564 [details]
Testing script for RHEL 7.6 / 7.7 / 7.8
Created attachment 1619630 [details] Output from the testing script TEST SETUP 4 Distro: RHEL-7.8-20190912.3 Server x86_64 Component version: NetworkManager-1.18.0-5.el7_7.1.x86_64 NetworkManager-ovs-1.18.0-5.el7_7.1.x86_64 TEST PROCEDURE 4 1. Create a Beaker job to reproduce the environment: https://beaker.engineering.redhat.com/jobs/3810651 2. Install Brew build NetworkManager-1.18.0-6.el7: https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=969104 3. Install Brew builds containing packages of Open vSwitch suitable for RHEL 7.8: - openvswitch2.11-2.11.0-23.el7fdn - openvswitch-selinux-extra-policy-1.0-13.el7fdp 4. To troubleshoot NetworkManager, follow the steps from: https://cgit.freedesktop.org/NetworkManager/NetworkManager/tree/contrib/fedora/rpm/NetworkManager.conf#n28 # sed -i "/RateLimitInterval/d" /etc/systemd/journald.conf # sed -i "/RateLimitBurst/d" /etc/systemd/journald.conf # echo -e "RateLimitInterval=0s\nRateLimitBurst=0" >> /etc/systemd/journald.conf # systemctl restart systemd-journald 5. Run the testing script. 6. Get messages from the system journal. RESULT 4 I do NOT get the expected result from that testing environment. Output of the script is attached. Created attachment 1619631 [details]
System journal from reproducing the bug under RHEL-7.8-20190912.3
Created attachment 1619633 [details]
NM trace log from reproducing the bug under RHEL-7.8-20190912.3
From the log: <info> [1569521361.6054] ovsdb: Could not connect: No such file or directory <debug> [1569521361.6054] ovsdb: disconnecting from ovsdb <warn> [1569521361.6055] device brcnv-iface could not be added to a ovs port: Request cancelled Is the openvswitch service running? (In reply to Beniamino Galvani from comment #32) > From the log: > > <info> [1569521361.6054] ovsdb: Could not connect: No such file or > directory > <debug> [1569521361.6054] ovsdb: disconnecting from ovsdb > <warn> [1569521361.6055] device brcnv-iface could not be added to a ovs > port: Request cancelled > > Is the openvswitch service running? [root@qe-dell-ovs5-vm-2 ~]# systemctl status openvswitch ● openvswitch.service - Open vSwitch Loaded: loaded (/usr/lib/systemd/system/openvswitch.service; disabled; vendor preset: disabled) Active: inactive (dead) Please retest after starting the service with: systemctl start openvswitch Created attachment 1619933 [details] Output from the testing script I enabled the openvswitch service. I used the same testing setup and procedure as written in comment #29. RESULT 5 The expected result has been achieved. Output of the script is attached. Created attachment 1619935 [details]
System journal from reproducing the bug under RHEL-7.8-20190912.3
Created attachment 1619937 [details]
NM trace log from reproducing the bug under RHEL-7.8-20190912.3
OUTCOME 1. Upgrading component: NetworkManager from: 1.12.0-10.el7_6 to: 1.18.0-6.el7 2. Installing packages from Brew server: - openvswitch2.11-2.11.0-23.el7fdn - openvswitch-selinux-extra-policy-1.0-13.el7fdp 3. Starting service openvswitch has fixed the bug. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:1162 |
Created attachment 1593891 [details] NetworkManager journal w/ 'trace' level logging Description of problem: We are using NetworkManager with the OVS plugin to configure a bridge on top of a bonded pair of interfaces. After restarting NetworkManager, only one slave interface is connected. All other ports/bridges/interfaces need to be manually restored with 'nmcli conn up ...' Version-Release number of selected component (if applicable): NetworkManager-ovs-1.12.0-10.el7_6.x86_64 NetworkManager-1.12.0-10.el7_6.x86_64 How reproducible: Always Steps to Reproduce: 1. Install openvswitch and the NetworkManager-ovs plugin 2. Configure bonds and a bridge as below: nmcli conn add type ovs-bridge conn.interface brcnv nmcli conn add type ovs-port conn.interface port0 master brcnv nmcli conn add type ovs-port conn.interface brcnv-port master brcnv nmcli conn add type ovs-interface conn.id brcnv-iface conn.interface brcnv master brcnv-port ipv4.method manual ipv4.address $ip ipv4.gateway $gw ipv4.dns "$dns" nmcli conn add type ovs-port conn.interface bond0 master brcnv ovs-port.bond-mode balance-slb nmcli conn add type ethernet conn.interface $default_device master bond0 nmcli conn add type ethernet conn.interface $secondary_device master bond0 nmcli conn down "System $default_device" nmcli conn mod "System $default_device" connection.autoconnect no 3. systemctl restart NetworkManager Actual results: # nmcli conn NAME UUID TYPE DEVICE ovs-slave-em3 9f8bf683-619d-40de-9174-d6d296733d80 ethernet em3 <green> <...snip / all others 'down' > Expected results: # nmcli conn NAME UUID TYPE DEVICE brcnv-iface cbfe1830-31ff-4a6b-ac52-86f5a1be60e7 ovs-interface brcnv ovs-bridge-brcnv 6c65d0db-e7fe-42b0-8b3c-8be1e5dc508e ovs-bridge brcnv ovs-slave-bond0 f3ec130f-2163-4093-8479-085a0ac815e4 ovs-port bond0 ovs-slave-brcnv-port 1f6e5b0a-5a02-4e06-a85c-c1c55f27e183 ovs-port brcnv-port ovs-slave-em1 b6badd81-f431-48cb-9916-5e1d73b5e1ad ethernet em1 ovs-slave-em3 cb51ba17-8e74-411b-a489-903606da4320 ethernet em3 ovs-slave-port0 aaf0ad18-0ad6-4a66-9cbf-ed6e3c6284c2 ovs-port port0 < ... snip ... > Additional info: Can restore configuration by: nmcli conn up ovs-bridge-brcnv nmcli conn up brcnv-iface nmcli conn up ovs-slave-em1