New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ipv6 has been disabled in server, but ISO boots server with ipv6 assigned to eth0 and lo #748
Comments
@blaxxton please run rear and capture the following output (of scripts). Perhaps use GitHub Gist to store the output.
|
@blaxxton perhaps also list the |
@blaxxton how did you disable ipv6 on your server? |
|
I deleted the ipv6 address with ifconfig command after booting to rescue by choosing recover. |
MyHostName###ls -la /etc/sysconfig/network-scripts/ifcfg-* |
ipv6 disabled on server by adding:
|
Did you define |
Already tired USE_STATIC_NETWORKING=1 in /etc/rear/local.conf MyHostName###tail -n 2 /etc/rear/local.conf /etc/rear/local.conf does not have any USE_DHCLIENT. I can't find where ipv6 is being set in any of the scripts. MyHostName###rpm -aq | grep -i dhcp . |
Correct me if I am wrong. In line 64 of 21_include_dhclient.sh: This means if DHCLIENT6_BIN is not empty set USE_DHCLIENT to y which seems to be correct MyHostName###export DHCLIENT6_BIN=""; echo $DHCLIENT6_BIN; [[ ! -z "$DHCLIENT6_BIN" ]] && USE_DHCLIENT=y; echo $USE_DHCLIENT y I am using bash. |
I even tried passing the ip and nm to kernel at boot time, but still ipv6 is being setup on both eth0 and loopback. |
MyHostName###export DHCLIENT6_BIN= ; DHCLIENT6_BIN=${DHCLIENT6_BIN##*/}; echo $DHCLIENT6_BIN; [ ! -z "$DHCLIENT6_BIN" ] && USE_DHCLIENT=y; echo $USE_DHCLIENT y |
I agree, I don't see any command that set ipv6 in rootfs, but I think some how USE_DHCLIENT is being set and dhcpd6 is being called. is it possible it is defined in kernel config file ? |
@blaxxton to be 100% sure you could add |
I don't think running set -x option to debug would help because you were right, and there is no where in script that ipv6 being set using ip command. adding EXCLUDE_MODULES=( ipv6 ) to /etc/rear/local.conf fixed my problem and not seeing any ipv6 address assigned to either loopback or eth0. This is RedHat 6 and even after disabling ipv6 in /etc/modprobe.d/ipv6.conf, I can see ipv6 modules in modprobe -l output. According to RedHat Document: The root cause of the issue is the dracut-network plug-in but we don't have such a plugin installed. I tried rebuilding the initramfs but still am able to see ipv6 modules also not a very good idea to disable ipv6 module at system level. somehow , your script create a initramfs that has ipv6 enabled based on the server. |
perhaps you could run once |
Only one entry in that file: but there is ipv6 module in /var/lib/rear/recovery/initrd_modules grep ipv6 var/lib/rear/recovery/initrd_modules ipv6 is being added to interfaces when kernel is being loaded. |
Interesting feedback - the script |
IMHO we should actively suppress IPv6 as long as we don't handle it properly (with a flag to suppress that mechanism). So far we had more reports of "IPv6 broke my environment" than of "Please handle IPv6 properly". |
Is a good starting point |
I think it is a good idea to disable ipv6 since ipv4 is always availablein all environments, even those that has set up ipv6 on their networkbut ipv6 would cause problem for those that has not setup ipv6 in thier environments and add extra steps to disable ipv6 to be able to recover.
To: rear/rear rear@noreply.github.com Interesting feedback - the script ./usr/share/rear/rescue/GNU/Linux/26_collect_initrd_modules.sh prepares the initrd_modules file. |
One more note: Thanks
To: rear/rear rear@noreply.github.com Interesting feedback - the script ./usr/share/rear/rescue/GNU/Linux/26_collect_initrd_modules.sh prepares the initrd_modules file. |
IMHO ReaR should faithfully copy what you did on your source system. If you A quick look reveals that sysctl is not handled by ReaR whatsoever. Maybe On 5 January 2016 at 05:56, blaxxton notifications@github.com wrote:
|
@blaxxton perhaps you give it a try with development version rear-1.17.2-git201601052030? |
I am going to try it tomorrow, will let you know the result soon.
To: rear/rear rear@noreply.github.com @blaxxton perhaps you give it a try with development version rear-1.17.2-git201601052030?— |
Can you please provide the link to download rear-1.17.2-git201601052030 ?
To: rear/rear rear@noreply.github.com @blaxxton perhaps you give it a try with development version rear-1.17.2-git201601052030?— |
See page http://relax-and-recover.org/download/ and select a snapshot version according your distro and go deeper into x86_64 directory. |
My test server is RedHat 6, but there is no x86_64 on snapshot version. Index of /repositories/Archiving:/Backup:/Rear:/Snapshot/RedHat_RHEL-6/i686 | | rear-1.17.2-115.git201601052030.el6.i686.rpm Is this ok ?
To: rear/rear rear@noreply.github.com See page http://relax-and-recover.org/download/ and select a snapshot version according your distro and go deeper into x86_64 directory.— |
@blaxxton yes that is ok |
I just tried rear-1.17.2-118.git201601091610.el6.i686.rpm or later error:2016-01-11 17:18:31.334417502 Saving result files with NSR (EMC NetWorker)2016-01-11 17:18:31.661689735 Saving files '/var/lib/rear/output/rear-clientname.iso' with save98519:save: Unable to setup direct save with server nsr_server: no matching IP interface data domain devices for save of client
To: rear/rear rear@noreply.github.com @blaxxton yes that is ok— |
@blaxxton could you check the |
Yes, files have been created: ls -latr /tmp/rear.CztH5P9ZiI8iI4L/rootfs/var/lib/rear/recoverytotal 40drwxr-xr-x 5 root root 4096 Jan 9 20:22 ../-rw-r--r-- 1 root root 143 Jan 12 08:36 mountpoint_permissions-rw-r--r-- 1 root root 665 Jan 12 08:36 diskbyid_mappings-rw-r--r-- 1 root root 5 Jan 12 08:36 bootloader-rw-r--r-- 1 root root 112 Jan 12 08:36 mountpoint_device-rw-r--r-- 1 root root 91 Jan 12 08:36 storage_drivers-rw-r--r-- 1 root root 11 Jan 12 08:36 nsr_server-rw-r--r-- 1 root root 2997 Jan 12 08:36 initrd_modules-r--r--r-- 1 root root 0 Jan 12 08:36 if_inet6drwxr-xr-x 2 root root 4096 Jan 12 08:36 ./-rw-r--r-- 1 root root 24 Jan 12 08:36 nsr_paths What I noticed, if_inet6 is with size of 0 either with older version or the current one. The actual error is complaining about: but old version can do the backup with no problem. |
@blaxxton Did you still had an IPv6 address after you booted from the rescue image? |
Still ipv6 there. I think becuase 65-sysctl.sh only including /etc/sysctl.d/*.conf and missing including /etc/sysctl.conf. run sysctl /etc/sysctl.conf and /etc/sysctl.d/*.conf filesfor file in /etc/sysctl.d/*.conf ; do nsr_server and nsr_paths seems to be ok and contain the server name and file systems are being backed up. From: gdha notifications@github.com @blaxxton Did you still had an IPv6 address after you booted from the rescue image? |
@blaxxton did you re-run your tests with the latest modifications? In rescue mode you could run |
Ok, thanks, the scripts is fixed and now it can read /etc/sysctl.conf, but still there is ipv6 address assigned after booting from ISO. Booted from ISO and sysctl -a | grep -i ipv6 | grep disable shows that ipv6 is enabled. Some how 65-sysctl.sh script is not being executed at boot time. Can you please check why this script is not being called or executed when we boot from ISO ? Thanks |
Can we disable backing up the ISO file in NSR ?We run rear mkrescue to create the ISO to boot from but at the end it backup the ISO file as well. |
@blaxxton Can we disable backing up the ISO file in NSR ? |
@blaxxton Some how 65-sysctl.sh script is not being executed at boot time. |
@blaxxton I've tested the |
If no-one objects I'll mark it as fixed |
Hi
To: rear/rear rear@noreply.github.com Closed #748.— |
ipv6 is disabled on server but when we boot the server from ISO and
choose Recover, there is ipv6 assigned to either eth0 and lo.
This cause problem recovering from EMC Legato Networker with ipv6 name is not accessible.
EMC legato Networker check if there is ipv6 assigned to interface, and if there is , will look to find
the server name based on ipv6 address defined in /etc/hosts and fails.
We had to delete the ipv6 address on both eth0 and loopback to be able to communicate with EMC Networker server and recover.
There is an option of adding ipv6 to /etc/hosts, but even with that , autorecover from Legato would not work if ipv6 is enabled in rescue mode while the original server has ipv6 disabled.
There should be an option to disable ipv6.
Thanks
The text was updated successfully, but these errors were encountered: