RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Bug 1245951 - Corosync cannot bind to IPv6 address
Summary: Corosync cannot bind to IPv6 address
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Linux 7
Classification: Red Hat
Component: initscripts
Version: 7.1
Hardware: Unspecified
OS: Unspecified
medium
medium
Target Milestone: rc
: ---
Assignee: initscripts Maintenance Team
QA Contact: Leos Pol
URL:
Whiteboard:
: 1324546 (view as bug list)
Depends On:
Blocks: 1324546
TreeView+ depends on / blocked
 
Reported: 2015-07-23 08:21 UTC by Sergey Safarov
Modified: 2018-07-05 06:39 UTC (History)
14 users (show)

Fixed In Version: initscripts-9.49.31-1.el7
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2016-11-04 06:42:15 UTC
Target Upstream Version:
Embargoed:

Attachments (Terms of Use)
system boot log (286.60 KB, text/plain)
2015-07-23 08:21 UTC, Sergey Safarov 		no flags Details
corosync ipv6 example (482 bytes, text/plain)
2015-07-23 08:27 UTC, Sergey Safarov 		no flags Details
NIC config files (652 bytes, application/x-gzip)
2016-02-16 14:53 UTC, Sergey Safarov 		no flags Details
View All


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 1334439 0 low CLOSED Unable to disable IPv6 DAD or Optimistic DAD for all interfaces 2021-02-22 00:41:40 UTC
Red Hat Bugzilla 1378917 0 medium CLOSED ifup waits for dadfailed tentative addresses 2021-02-22 00:41:40 UTC
Red Hat Product Errata RHBA-2016:2456 0 normal SHIPPED_LIVE initscripts bug fix update 2016-11-03 14:04:56 UTC

Internal Links: 1334439 1378917

Description Sergey Safarov 2015-07-23 08:21:11 UTC 
Created attachment 1055209 [details]
system boot log

Corosync cannot automaticaly start after server reboot
Used corosync-2.3.4-4


To reproducible:
1. Configure IPv6 only address
2. Configure corosync to bind IPv6 addess (example in attached file)
3. Reboot server
4. Login and sed corosync.service status, system journal

Actual results:
[root@couch1 ~]# journalctl -f 
-- Logs begin at Чт 2015-07-23 08:10:23 GMT. --
июл 23 08:10:40 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2783-41)
июл 23 08:10:40 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2785-42)
июл 23 08:10:41 couch1.db.rcsnet.ru corosync[1388]: [MAIN  ] Totem is unable to form a cluster because of an operating system or network ...operly.
июл 23 08:10:41 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2787-43)
июл 23 08:10:41 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2789-44)
июл 23 08:10:42 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2791-45)
июл 23 08:10:42 couch1.db.rcsnet.ru corosync[1388]: [MAIN  ] Totem is unable to form a cluster because of an operating system or network ...operly.
июл 23 08:10:42 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2793-46)
июл 23 08:10:43 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2795-47)
июл 23 08:10:43 couch1.db.rcsnet.ru corosync[1388]: [QB    ] Denied connection, is not ready (1388-2797-48)


[root@couch1 ~]# systemctl status corosync.service
corosync.service - Corosync Cluster Engine
   Loaded: loaded (/usr/lib/systemd/system/corosync.service; enabled)
   Active: failed (Result: exit-code) since Чт 2015-07-23 08:11:26 GMT; 17s ago
  Process: 1283 ExecStart=/usr/share/corosync/corosync start (code=exited, status=1/FAILURE)

июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [QB    ] withdrawing server sockets
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [SERV  ] Service engine unloaded: corosync configuration map access
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [QB    ] withdrawing server sockets
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [SERV  ] Service engine unloaded: corosync configuration service
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [QB    ] withdrawing server sockets
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [SERV  ] Service engine unloaded: corosync cluster closed process group service v1.01
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [QB    ] withdrawing server sockets
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [SERV  ] Service engine unloaded: corosync cluster quorum service v0.1
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [SERV  ] Service engine unloaded: corosync profile loading service
июл 23 08:11:26 couch1.db.rcsnet.ru corosync[1388]: [MAIN  ] Corosync Cluster Engine exiting normally

[root@couch1 ~]# netstat -anp --inet6 | grep coro
Empty output

Expected results:
[root@couch1 ~]# netstat -anp --inet6 | grep coro
udp6       0      0 fd75:9fac:a7b1:92:35952 :::*                                2987/corosync       
udp6       0      0 fd75:9fac:a7b1:92:33911 :::*                                2987/corosync       
udp6       0      0 fd75:9fac:a7b1:92::1234 :::*                                2987/corosync       



Additional info:
In system log you can find 

corosync[1407]: [TOTEM ] bind token socket failed: Cannot assign requested address (99)
corosync[1407]: [TOTEM ] The network interface [fd75:9fac:a7b1:92:6c::901] is now up.
corosync[1407]: [SERV  ] Service engine loaded: corosync configuration map access [0]
corosync[1407]: [QB    ] server name: cmap
corosync[1407]: [SERV  ] Service engine loaded: corosync configuration service [1]
corosync[1407]: [QB    ] server name: cfg
corosync[1407]: [SERV  ] Service engine loaded: corosync cluster closed process group service v1.01 [2]
corosync[1407]: [QB    ] server name: cpg
corosync[1407]: [SERV  ] Service engine loaded: corosync profile loading service [4]
corosync[1407]: [QUORUM] Using quorum provider corosync_votequorum
corosync[1407]: [SERV  ] Service engine loaded: corosync vote quorum service v1.0 [5]
corosync[1407]: [QB    ] server name: votequorum
corosync[1407]: [SERV  ] Service engine loaded: corosync cluster quorum service v0.1 [3]
corosync[1407]: [QB    ] server name: quorum
corosync[1407]: [TOTEM ] adding new UDPU member {fd75:9fac:a7b1:92:6c::901}
corosync[1407]: [TOTEM ] bind token socket failed: Cannot assign requested address (99)
corosync[1407]: [TOTEM ] adding new UDPU member {fd75:9fac:a7b1:92:6c::902}
corosync[1407]: [TOTEM ] bind token socket failed: Cannot assign requested address (99)
Comment 1 Sergey Safarov 2015-07-23 08:27:32 UTC 
Created attachment 1055226 [details]
corosync ipv6 example
Comment 2 Sergey Safarov 2015-07-23 08:29:36 UTC 
To workaround add "ExecStartPre=/bin/sleep 3" to systemd unit file
Comment 6 Jan Friesse 2016-01-25 16:19:28 UTC 
I strongly believe this bug is caused by another component (probably NetworkManager, more specifically wait-nm-online).

Does problem persist in 7.2?
Comment 7 Sergey Safarov 2016-02-11 13:50:09 UTC 
I has tested on 7.2.1511 - issue is reproduced.
I has added second NIC to server with IPv4 address and error is resolved.

Network Manager is not installed on system

[root ~]# rpm -qa | grep NetworkManager
[root ~]#
Comment 8 Jan Friesse 2016-02-12 15:37:01 UTC 
Sergey,
ok, good news you resolved problem, but it may still make sense to solve main problem.

Actually, corosync is unable to wait itself to ip address. It's something we are planning (in upstream) for corosync 3.0.

But currently, corosync service file is using Requires=network-online.target, so it shouldn't be executed before all interfaces are up and running with all addresses assigned.

I'm moving bug to initscript (I'm not 100% sure about component, but it's not NetworkManager problem, because user is not using it).
Comment 9 Lukáš Nykrýn 2016-02-12 15:55:43 UTC 
Can you attach your ifcfg files?
Comment 10 Sergey Safarov 2016-02-16 14:53:31 UTC 
Created attachment 1127610 [details]
NIC config files
Comment 15 Lukáš Nykrýn 2016-04-25 10:14:36 UTC 
We don't have any explicit Duplicate Address Detection for ipv6 in intscripts and we don't touch accept_dad setting.

Everything that initscripts do is calling /sbin/ip -6 addr add $address dev $device
so lets try iproute.
Comment 17 Phil Sutter 2016-04-25 10:47:42 UTC 
(In reply to Lukáš Nykrýn from comment #15)
> We don't have any explicit Duplicate Address Detection for ipv6 in
> intscripts and we don't touch accept_dad setting.
> 
> Everything that initscripts do is calling /sbin/ip -6 addr add $address dev
> $device
> so lets try iproute.

I'm not sure this is something which should be fixed in iproute. Basically DAD is a kernel thing, and the default behaviour is to keep an address in tentative state until DAD is completed. While one could enhance iproute to optionally wait for an assigned address to become permanent, the same could be taken care of by initscripts, see here[1] for an example.

Another alternative might be to use Optimistic DAD[2] in RHEL by default. Not sure if a change at this scale is acceptable for a regular y-stream release though.


[1] https://www.vaspects.com/2013/12/11/services-dont-bind-to-ipv6-address/
[2] https://lwn.net/Articles/218597/
Comment 19 Lukáš Nykrýn 2016-04-25 15:37:59 UTC 
It would be nice to have an option for the wait in ip, but I would be fine even with the patch in initscripts (thanks for the link!).
Comment 21 Sergey Safarov 2016-05-06 08:48:27 UTC 
What type of info is requred?
Comment 22 Michele Baldessari 2016-05-06 09:00:07 UTC 
ops had a private comment by mistake.

Lukas,

are you planning on taking a stab at adding the wait in the initscripts
or should we pursue the kernel folks to see if they can fix the
net.ipv6.conf.{all,default}.{optimistic_dad,accept_dad} sysctl? (I confirmed
with Hannes that it they are not working ATM)

thanks,
Michele
Comment 23 Lukáš Nykrýn 2016-05-09 10:35:22 UTC 
Maybe both? I will reassign this bug back to initscript and add the wait.
Comment 25 Lukáš Nykrýn 2016-06-13 14:45:04 UTC 
https://git.fedorahosted.org/cgit/initscripts.git/commit/?h=rhel7-branch&id=8121b7f765e33e7d84eecd762e1c0d37964bfcaf -> post
Comment 27 Leos Pol 2016-09-23 13:49:30 UTC 
Tested with initscripts-9.49.37-1.el7:

During iface initialization I can see following output while address is in tentative state:
INFO     : [ipv6_wait_tentative] Waiting for interface eth0 IPv6 address(es) to leave the "tentative" state

I'm getting this each second for 30 sec at max, which corresponds to applied patch. This should fix reported problem.

Unfortunately when dad fails, address remains tentative and script remains in the loop for 30s.

# ip -6 addr show dev eth0 scope global tentative
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1400 state UP qlen 1000
    inet6 aaaa:bbbb:cccc::1234/64 scope global tentative dadfailed 
       valid_lft forever preferred_lft forever

At this stage, I rather call this as z-stream candidate. Do you agree, Lukas?
Comment 30 errata-xmlrpc 2016-11-04 06:42:15 UTC 
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://rhn.redhat.com/errata/RHBA-2016-2456.html
Comment 31 Fabio Massimo Di Nitto 2017-02-22 12:08:33 UTC 
*** Bug 1324546 has been marked as a duplicate of this bug. ***
Comment 32 Vijay 2018-07-05 06:39:36 UTC 
This issue is reproducible in initscripts-9.49.39-1.el7.x86_64 version on RHEL7.4
Note You need to log in before you can comment on or make changes to this bug.