Bug 1553625
| Summary: | ifcfg files of bond+vlan device are incorrect when configured via network cmd in kickstart file | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | jiachen zhang <jiaczhan> | ||||||
| Component: | anaconda | Assignee: | Ryan Barry <rbarry> | ||||||
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Release Test Team <release-test-team> | ||||||
| Severity: | urgent | Docs Contact: | |||||||
| Priority: | urgent | ||||||||
| Version: | 7.5 | CC: | asabadra, atragler, cshao, deekej, dfediuck, huzhao, jstodola, mtessun, qiyuan, rbarry, rvykydal, salmy, sbonazzo, sbueno, spower, weiwang, yaniwang, ycui, yturgema, yzhao | ||||||
| Target Milestone: | pre-dev-freeze | ||||||||
| Target Release: | --- | ||||||||
| Hardware: | Unspecified | ||||||||
| OS: | Unspecified | ||||||||
| Whiteboard: | |||||||||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |||||||
| Doc Text: | Story Points: | --- | |||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2018-07-16 15:05:25 UTC | Type: | Bug | ||||||
| Regression: | --- | Mount Type: | --- | ||||||
| Documentation: | --- | CRM: | |||||||
| Verified Versions: | Category: | --- | |||||||
| oVirt Team: | Node | RHEL 7.3 requirements from Atomic Host: | |||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||
| Embargoed: | |||||||||
| Bug Depends On: | |||||||||
| Bug Blocks: | 1447254 | ||||||||
| Attachments: |
|
||||||||
Isn't this a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1553595 ? It's not a duplicate. Although two bugs appeared when creating bond+vlan, the phenomenon is not the same. Ryan can we have a workaround in RHV-H while platform fixes this? Vlan device can get IP after restar NetworkManager. But after reboot host, vlan device still can't get IP. (In reply to cshao from comment #5) > Vlan device can get IP after restar NetworkManager. But after reboot host, > vlan device still can't get IP. You may mix up this bug with Bug 1553595, for this one, there is no vlan device, although bond0.50 is shown, it's a bond device actually. Restart NetworkManager cannot change anything. No workaround, and like the other, I am not sure one is possible The same issues occurs in redhat-virtualization-host-4.1-20180307.0 The suspicious thing here is that ifcfg files for the network command are not created in initramfs (or not transferred to installer root). Anaconda is not able to create all the required files correctly in stage2 then. The situation is equivalent to defining the networking in %pre section (where ifcfg files are not created in initramfs as well). I think for this case (vlan over bond) we are not able to deal with the configuration. What might help would be to try to create the devices in two steps: network --device=bond0 --noipv4 --bondslaves=em3,em4 --bondopts=mode=active-backup,primary=em3,miimon=100 network --device=bond0 --bootproto=dhcp --vlanid=50 but I am not sure if it actually works. To investigate more I'd need a link to RHEV iso I could use to reproduce the issue. Also the complete kickstart used for installation could be useful. Could you try to reproduce with rd.debug installer boot option added and attach the logs? (In reply to Radek Vykydal from comment #15) > Could you try to reproduce with rd.debug installer boot option added and > attach the logs? No more needed. I've reproduced the issue with RHVH-4.2-20180305.0-RHVH-x86_64-dvd1.iso locally. It seems that importing of the state (ifcfg files in this case) from initramfs is broken. (In reply to Radek Vykydal from comment #16) > No more needed. I've reproduced the issue with > RHVH-4.2-20180305.0-RHVH-x86_64-dvd1.iso locally. > It seems that importing of the state (ifcfg files in this case) from > initramfs is broken. rhel-import-state.service is not enabled which reminds me of bug #1537977 on Fedora. (In reply to jiachen zhang from comment #8) > The same issues occurs in redhat-virtualization-host-4.1-20180307.0 On http://download.englab.brq.redhat.com/released/RHV-4.1/RHVH-builds/RHVH-4.1-20180315.0/compose/RHVH/x86_64/os/ the service is enabled so I am surprised that you are seeing the problem here. Could you please provide anaconda logs for this case? There may be some additional problem and need to handle that as well. Created attachment 1414020 [details]
/var/log for rhvh-4.1-20180307.0
So sorry for deleteing the needinfo accidentally. So sorry for deleteing the needinfo accidentally. /usr/lib/systemd/system-preset/90-default.preset used for RHEV 4.2 doesn't seem to have the update from bug 1493545 which went into redhat-release-server package. Maybe the file in redhat-virtualization-host-productimg needs update? (In reply to jiachen zhang from comment #19) > Created attachment 1414020 [details] > /var/log for rhvh-4.1-20180307.0 Thank you. The problem here is the same as with 4.2 - rhel-import-state service not enabled. (In reply to Radek Vykydal from comment #10) > The suspicious thing here is that ifcfg files for the network command are > not created in initramfs (or not transferred to installer root). > > Anaconda is not able to create all the required files correctly in stage2 > then. The situation is equivalent to defining the networking in %pre section > (where ifcfg files are not created in initramfs as well). I think for this > case (vlan over bond) we are not able to deal with the configuration. What > might help would be to try to create the devices in two steps: > > network --device=bond0 --noipv4 --bondslaves=em3,em4 > --bondopts=mode=active-backup,primary=em3,miimon=100 > network --device=bond0 --bootproto=dhcp --vlanid=50 > > but I am not sure if it actually works. > > To investigate more I'd need a link to RHEV iso I could use to reproduce the > issue. Also the complete kickstart used for installation could be useful. I tried this in redhat-virtualization-host-4.2-20180326.0 and add "--noipv6" in the first command "network --device=bond0 --noipv4 --noipv6 --bondslaves=em3,em4 --bondopts=mode=active-backup,primary=em3,miimon=100". Actually, this time I can get bond0,bond0.50 and their ifcfg files, but still cannot get the Vlan ip. So I don't think this method works. So we need to add all of the following presets to our productimg (or just rhel-import-state) ? # https://bugzilla.redhat.com/show_bug.cgi?id=1493545 enable rhel-autorelabel.service enable rhel-autorelable-mark.service enable rhel-configure.service enable rhel-dmesg.service enable rhel-import-state.service enable rhel-loadmodules.service enable rhel-readonly.service I guess all of them (or those in https://bugzilla.redhat.com/show_bug.cgi?id=1493545#c15, ie containing also rhel-domainname.service - this is what RHEL7.5 has) but I'd rather ask dkaspar who handled the 1493545. Also accidentally cleared needinfo. Also accidentally cleared needinfo. For RHVH-4.1-20180331.1-RHVH-x86_64-dvd1.iso, when configure bond+vlan in kickstart file, the bond+vlan device and ifcfg files are normal now. |
Created attachment 1406175 [details] /var/log /tmp network-scripts Description of problem: Configure bond+vlan device in kickstart file as: network --device=bond0 --bootproto=dhcp --bondslaves=em3,em4 --bondopts=mode=active-backup,primary=em3,miimon=100 --vlanid=50 After installation finished, there is no bond0 device, bond0.50 is there, but without IP. [root@test ~]# ip addr 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: em1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether 24:6e:96:19:bb:70 brd ff:ff:ff:ff:ff:ff inet 10.73.131.63/23 brd 10.73.131.255 scope global noprefixroute dynamic em1 valid_lft 43077sec preferred_lft 43077sec inet6 2620:52:0:4982:f9b6:560f:188f:6ee5/64 scope global noprefixroute dynamic valid_lft 2591878sec preferred_lft 604678sec inet6 fe80::fa95:b7cc:aeb1:b8a8/64 scope link noprefixroute valid_lft forever preferred_lft forever 3: em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether 24:6e:96:19:bb:71 brd ff:ff:ff:ff:ff:ff 4: p5p1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether a0:36:9f:ae:9f:54 brd ff:ff:ff:ff:ff:ff 5: em3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether 24:6e:96:19:bb:72 brd ff:ff:ff:ff:ff:ff 6: em4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether 24:6e:96:19:bb:73 brd ff:ff:ff:ff:ff:ff 7: p7p1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000 link/ether a0:36:9f:9d:37:12 brd ff:ff:ff:ff:ff:ff 8: p7p2: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000 link/ether a0:36:9f:9d:37:13 brd ff:ff:ff:ff:ff:ff 9: p5p2: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000 link/ether a0:36:9f:ae:9f:56 brd ff:ff:ff:ff:ff:ff 20: bond0.50: <NO-CARRIER,BROADCAST,MULTICAST,MASTER,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000 link/ether 5e:07:3c:86:14:b4 brd ff:ff:ff:ff:ff:ff The ifcfg files are: [root@test network-scripts]# ll total 292 -rw-r--r--. 1 root root 390 Mar 9 15:19 ifcfg-bond0.50 -rw-r--r--. 1 root root 127 Mar 9 15:19 ifcfg-bond0.50_slave_1 -rw-r--r--. 1 root root 127 Mar 9 15:19 ifcfg-bond0.50_slave_2 -rw-r--r--. 1 root root 306 Mar 9 15:19 ifcfg-em1 -rw-r--r--. 1 root root 275 Mar 9 15:19 ifcfg-em2 -rw-r--r--. 1 root root 254 Jan 3 00:29 ifcfg-lo -rw-r--r--. 1 root root 277 Mar 9 15:19 ifcfg-p5p1 -rw-r--r--. 1 root root 277 Mar 9 15:19 ifcfg-p5p2 -rw-r--r--. 1 root root 277 Mar 9 15:19 ifcfg-p7p1 -rw-r--r--. 1 root root 277 Mar 9 15:19 ifcfg-p7p2 [root@test network-scripts]# cat ifcfg-bond0.50 BONDING_OPTS="miimon=100 mode=active-backup primary=em3" TYPE="Bond" BONDING_MASTER="yes" PROXY_METHOD="none" BROWSER_ONLY="no" BOOTPROTO="dhcp" DEFROUTE="yes" IPV4_FAILURE_FATAL="no" IPV6INIT="yes" IPV6_AUTOCONF="yes" IPV6_DEFROUTE="yes" IPV6_FAILURE_FATAL="no" IPV6_ADDR_GEN_MODE="stable-privacy" NAME="bond0.50" UUID="c8ced0e7-0aab-45b5-a8fe-a88899b3643a" DEVICE="bond0.50" ONBOOT="yes" [root@test network-scripts]# cat ifcfg-bond0.50_slave_1 TYPE=Ethernet NAME="bond0.50 slave 1" UUID=842c9042-71df-4b22-a8cc-be9ee7491834 DEVICE=em3 ONBOOT=no MASTER=bond0.50 SLAVE=yes It seems ifcfg-bond0.50, which should be the configuration file of the vlan device, is the configuration file of bond0. And there is no ifcfg file for vlan device. Version-Release number of selected component (if applicable): RHVH-4.2-20180305.0-RHVH-x86_64-dvd1.iso How reproducible: 100% Steps: 1. Create a kickstart file with bond+vlan configuration as above, and other neccessary configurations. 2. After installation finished, check IP and ifcfg files. Actual Results: 1. No IP with the shown vlan device. 2. Missing ifcfg-bond0 file 3. ifcfg-bond0.50 is the configuration of bond device, not vlan device. Expected Results: 1. Vlan device can get IP. 2. ifcfg files for bond and vlan device should be correct.