Linux - NetworkingThis forum is for any issue related to networks or networking.
Routing, network cards, OSI, etc. Anything is fair game.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Hi, I hope you can help because I worked with my partner today and he is at a loss on this and I can't explain it
either. I will describe it the best I can since he set this up. We setup what he called 4 port bonding failover. The 4 ports were defined as slaves and bond0 and bond1 were setup. He defined bond0.conf and bond1.conf in the /etc/modprobe.d directory. Below are the configuration files.
The interfaces are up...eth1, eth2, eth3, eth4 and the network 'seems to function' but
it's not quite right, something seems wrong or slow, and there seem to be indications of excessive slowness for possibly a second sometimes.
Let me know if you can help it would be greatly appreciated.
Below are the file configurations made the only other change was entering the command -->> modprobe bonding
In your listing of /etc/sysconfig/network-scripts
Your first result is from cat-ing a backup file
You don't show the result of cat-ing the real file (eth0)
But in that backup (ifcfg-eth0~) you have MASTER=bondo not MASTER=bond0
If the real file is the same, that might be a problem.
Are you using gEdit ? You can disable those backups.
This ping is from a machine on the same switch.
Is this good response time?
[root@hhlxnrcsas02 /]# ping hhlxnrcsas01
PING hhlxnrcsas01.hh.chs (10.35.13.66) 56(84) bytes of data.
64 bytes from 10.35.13.66: icmp_seq=2 ttl=64 time=0.129 ms
64 bytes from 10.35.13.66: icmp_seq=5 ttl=64 time=0.116 ms
64 bytes from 10.35.13.66: icmp_seq=7 ttl=64 time=0.113 ms
64 bytes from 10.35.13.66: icmp_seq=9 ttl=64 time=0.106 ms
64 bytes from 10.35.13.66: icmp_seq=11 ttl=64 time=0.112 ms
64 bytes from 10.35.13.66: icmp_seq=13 ttl=64 time=0.118 ms
64 bytes from 10.35.13.66: icmp_seq=15 ttl=64 time=0.108 ms
64 bytes from 10.35.13.66: icmp_seq=17 ttl=64 time=0.109 ms
64 bytes from 10.35.13.66: icmp_seq=19 ttl=64 time=0.090 ms
64 bytes from 10.35.13.66: icmp_seq=21 ttl=64 time=0.092 ms
64 bytes from 10.35.13.66: icmp_seq=36 ttl=64 time=0.112 ms
64 bytes from 10.35.13.66: icmp_seq=37 ttl=64 time=0.107 ms
64 bytes from 10.35.13.66: icmp_seq=38 ttl=64 time=0.105 ms
64 bytes from 10.35.13.66: icmp_seq=39 ttl=64 time=0.119 ms
64 bytes from 10.35.13.66: icmp_seq=40 ttl=64 time=0.104 ms
64 bytes from 10.35.13.66: icmp_seq=60 ttl=64 time=0.113 ms
64 bytes from 10.35.13.66: icmp_seq=62 ttl=64 time=0.123 ms
64 bytes from 10.35.13.66: icmp_seq=64 ttl=64 time=0.115 ms
64 bytes from 10.35.13.66: icmp_seq=66 ttl=64 time=0.106 ms
64 bytes from 10.35.13.66: icmp_seq=68 ttl=64 time=0.093 ms
64 bytes from 10.35.13.66: icmp_seq=70 ttl=64 time=0.114 ms
64 bytes from 10.35.13.66: icmp_seq=72 ttl=64 time=0.113 ms
64 bytes from 10.35.13.66: icmp_seq=74 ttl=64 time=0.105 ms
64 bytes from 10.35.13.66: icmp_seq=76 ttl=64 time=0.115 ms
RE:May I ask, in your ping output, where is responses to icmp_seq=1,3,4,6,8,10,12....
Nice spot yes there is 50% packet loss below.
I must reiterate, I did not setup this up my partner did and I understand the basics of bonding, but not the configuration of bonding failover. Does the configuration look correct with the fixes to the typos? I will check out the source forge site as well.
Should I bail on this bonding failover idea since I have to get this machine networked for all the next phases of the project.
Below is the packet loss.
Below is a ping from a good machine on the same switch to the bonding failover machine.
[root@hclxnrcas01 ~]# ping hhlxnrcsas02
PING hhlxnrcsas02.hh.chs (10.35.13.85) 56(84) bytes of data.
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=2 ttl=64 time=1.16 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=4 ttl=64 time=0.087 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=6 ttl=64 time=0.084 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=9 ttl=64 time=0.104 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=11 ttl=64 time=0.130 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=13 ttl=64 time=0.131 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=15 ttl=64 time=0.129 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=17 ttl=64 time=0.129 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=19 ttl=64 time=0.127 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=21 ttl=64 time=0.128 ms
64 bytes from hhlxnrcsas02.hh.chs (10.35.13.85): icmp_seq=23 ttl=64 time=0.134 ms
You say you fixed the typos in the ifcfg-bondX files, does that mean they look like :
Code:
alias bond1 bonding
options bond1 -o bond1 mode=5 miimon=100
or do they still look like
Code:
alias bond0 bonding
alias bond0 mode=5 miimon=100
Why two alias lines ? Surely the second should specify the options ?
It's hard to track down problems when the basic outlines for bonding configuration have not been followed. Notice my links are specific fedora documentation for Fedora 12.
RE: I'm concerned that the proscribed way of creating ifcfg-bondX files is different
To be clear, I completely relied on my partner to configure this, so I can't answer as to why he decided on certain configurations.
Here is what he stated the goal was (as best I understand):
- "create bonding failover" using the 4 ports, with 2 bonds
- Physically the machine has (4) ports
- we created a dns entry for hhlxnrcsas02 - 10.35.13.85 for the primary
bond0 (eth1, eth2)
- we created a dns entry for hhlxnrcsas02b - 10.35.13.86 for the failover bond1 (eth3, eth4)
- he mentioned being uncomfortable with a fedora 12 deployment and ext4
since he thinks ext4 is not tested enough...but I assured him I researched this and ext4 is stable enough for this (2 server 120TB deployment - we are viewing hhlxnrcsas02 - one server)
He also said "there is no /etc/modprobe.conf but there is a /etc/modprobe.d directory" and below are the contents:root@hhlxnrcsas02 etc]# cd modprobe.d
[root@hhlxnrcsas02 modprobe.d]# ls -ltr
total 52
-rw-r--r--. 1 root root 16 2008-01-04 03:43 blacklist-visor.conf
-rw-r--r--. 1 root root 30 2009-07-07 00:21 openfwwf.conf
-rw-r--r--. 1 root root 845 2009-07-26 07:04 blacklist.conf
-rw-r--r--. 1 root root 473 2009-10-21 00:25 dist-oss.conf
-rw-r--r--. 1 root root 5616 2009-10-21 00:25 dist.conf
-rw-r--r--. 1 root root 382 2009-10-21 00:25 dist-alsa.conf
-rw-r--r--. 1 root root 53 2009-11-04 19:24 floppy-pnp.conf
-rw-r--r--. 1 root root 52 2010-02-26 08:50 anaconda.conf
-rw-r--r--. 1 root root 22 2010-02-26 12:12 bond0.conf~
-rw-r--r--. 1 root root 20 2010-02-26 12:15 bond1.conf~
-rw-r--r--. 1 root root 52 2010-02-26 12:34 bond0.conf
-rw-r--r--. 1 root root 50 2010-02-26 21:25 bond1.conf
[root@hhlxnrcsas02 modprobe.d]# cd ..
[root@hhlxnrcsas02 etc]# ls -ltr modprobe.conf
ls: cannot access modprobe.conf: No such file or directory
[root@hhlxnrcsas02 etc]#
So he put (what I think) would normally go in the /etc/modprobe.conf
file and created these bond0.conf and bond1.conf files with the entries
below since he stated there "was no /etc/modprobe.conf" He then ran
the command 'modprobe bonding'
Just an FYI I also saw him run the command '/sbin/insmod bond0'
and from what I read you need options for this and this is for tuning.
As you can see there is no /etc/modprobe.conf file so I cannot speak about this, should there be? (I'll keep reading...)
As far as the typo goes I corrected from your previous response:
[root@hhlxnrcsas02 modprobe.d]# cat bond1.conf
alias bond1 bonding
alias bond0 mode=5 miimon=100
[root@hhlxnrcsas02 modprobe.d]#
- NO - I have not changed the 2nd line to options but it looks like I
should according to what I have learned so far and I think I should?
(I'll keep reading the docs...)
- Is mode 5 correct for the goal we are trying to achieve?
This is the whole picture...I'll keep reading...
Let me know what you think...thank you
Personally, I like to keep things simple and "by the book" for one reason really. If something works, but there are errors or inconsistencies, it is very hard to trace those problems if you have strayed from the usual path. Sometimes things work even when it has been set up wrongly, it's not always black or white.
I accept that modprobe.conf is missing, and cannot comment on that. Maybe a separate googling will establish why that is.
Regarding the contents of the bondX.conf files, alias is a command.
If you use it as a label (which appears to be the case in line 2 of those files) you are actually aliasing bondX again, but with unpredictable results. The syntax is not correct for either alias or for assigning options.
This is as far as I can establish. I can find no other evidence of it being done like that.
And in bond1.conf, there should be no refs to bond0
your last post shows it as it was when I mentioned it before.
Another interesting thing is you say there are 4 physical ports, but you have listings for eth0, eth1, eth2, eth3, eth4.
Just in case, can you check with "ifconfig -a" and when just "ifconfig", what all ethernet card are active and have the same configuration. Also all of them connected to the same equipment.
RE: Personally, I like to keep things simple
RE: ifconfig....
-->Are you stating this "bonding failover" using bond0 and bond1 was a bad idea and I should revert to a 4x1 ? (I'll keep reading...)
- I'll find out what I can about the fed. 12 missing /etc/modprobe.conf
- I have just changed the bond0 and bond1.conf files to as follows:
changes made /etc/modprobe.d/bond0.conf
alias bond0 bonding
options bond0 -o bond0 mode=5 miimon=100
changes made /etc/modprobe.d/bond1.conf
alias bond1 bonding
options bond1 -o bond1 mode=5 miimon=100
(I followed your suggestion and need to check the syntax of: options bond1 -o bond1 mode=5 miimon=100)
- "And in bond1.conf, there should be no refs to bond0
your last post shows it as it was when I mentioned it before." ----> I have corrected this a few posts ago
- "Another interesting thing is you say there are 4 physical ports, but you have listings for eth0, eth1, eth2, eth3, eth4."
---> Actually there is more interfaces than needed...good spot...What should ifcfg-eth(x) should I start with (I'll keep reading)?????
Below is another clue to this problem and may be causing the packet loss which did not disappear after change the 2nd line in the bond0.conf and bond1.conf
files and issuing-->> 'modprobe bonding' and 'service network restart' commands.
-->TOO MANY interfaces per below should I start at ifcfg-eth0 or ifcfg-eth1?? --->see below
Thanks.
AFTER CHANGE DETAILED BELOW ---
changes made /etc/modprobe.d/bond0.conf
alias bond0 bonding
options bond0 -o bond0 mode=5 miimon=100
alias bond1 bonding
options bond1 -o bond1 mode=5 miimon=100
~
still plenty of packet loss below
[root@hhlxnrcsas02 modprobe.d]# ping hhlxnrcsas01
PING hhlxnrcsas01.hh.chs (10.35.13.66) 56(84) bytes of data.
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=3 ttl=64 time=0.122 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=6 ttl=64 time=0.116 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=8 ttl=64 time=0.115 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=10 ttl=64 time=0.104 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=12 ttl=64 time=0.107 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=14 ttl=64 time=0.115 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=16 ttl=64 time=0.114 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=18 ttl=64 time=0.112 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=20 ttl=64 time=0.112 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=22 ttl=64 time=0.109 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=24 ttl=64 time=0.109 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=26 ttl=64 time=0.117 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=28 ttl=64 time=0.108 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=29 ttl=64 time=0.117 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=30 ttl=64 time=0.107 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=31 ttl=64 time=0.106 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=117 ttl=64 time=0.107 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=119 ttl=64 time=0.083 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=121 ttl=64 time=0.112 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=123 ttl=64 time=0.116 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=125 ttl=64 time=0.120 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=127 ttl=64 time=0.116 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=129 ttl=64 time=0.114 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=131 ttl=64 time=0.114 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=133 ttl=64 time=0.117 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=135 ttl=64 time=0.116 ms
64 bytes from hhlxnrcsas01.hh.chs (10.35.13.66): icmp_seq=137 ttl=64 time=0.117 ms
^C
--- hhlxnrcsas01.hh.chs ping statistics ---
138 packets transmitted, 27 received, 80% packet loss, time 141998ms
rtt min/avg/max/mdev = 0.083/0.111/0.122/0.016 ms
[root@hhlxnrcsas02 modprobe.d]#
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.
