I am trying to find out why the interconnect between the nodes of the REDFERN cluster does not support jumbo frames as described in issue #47 (“Ping tests show connectivity across interconnect but large packets are dropped”).
I am trying to find why the configuration settings are not being applied.
Run the following command to change to the directory where the NIC configurations files are kept:
cd /etc/sysconfig/network-scripts
Run the following command to find what MTU settings are defined for the NICs:
grep -i mtu= ifcfg*
The expected output is:
ifcfg-eth0:MTU="9000" ifcfg-eth1:MTU="9000" ifcfg-eth1:1:MTU="9000" ifcfg-eth1:2:MTU="9000" ifcfg-eth1:3:MTU="9000" ifcfg-eth2:MTU="9000" ifcfg-xenbr0:MTU=9000 ifcfg-xenbr1:MTU="9000" ifcfg-xenbr2:MTU="9000" ifcfg-xenbr3:MTU="9000"
All bridges (xenbr*) have the MTU set to 9000.
Run the following command to see what the actual NIC settings are:
ip link show
The expected output is:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc pfifo_fast master xenbr0 state UP qlen 1000 link/ether 44:37:e6:a7:34:9a brd ff:ff:ff:ff:ff:ff 3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc pfifo_fast state UP qlen 1000 link/ether 00:1b:21:d9:a1:ae brd ff:ff:ff:ff:ff:ff 4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN qlen 1000 link/ether 00:1b:21:d9:a1:af brd ff:ff:ff:ff:ff:ff 5: xenbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP link/ether 44:37:e6:a7:34:9a brd ff:ff:ff:ff:ff:ff 6: xenbr1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN link/ether 82:84:44:75:b6:bf brd ff:ff:ff:ff:ff:ff 7: xenbr2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN link/ether 8a:ce:1e:a5:7a:d3 brd ff:ff:ff:ff:ff:ff 8: xenbr3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN link/ether 86:6f:7c:ea:a3:b2 brd ff:ff:ff:ff:ff:ff
Only the first bridge (xenbr0) has the MTU set to 9000. All other bridges have MTU set to 1500 (the default).
Run the following command to see if I set the MTU directly on the second bridge (xenbr1):
ip link set xenbr1 mtu 9000
The expected output is:
RTNETLINK answers: Invalid argument
See “Oracle VM: Change MTU Value Failed with 'Invalid argument' Running Kernel UEK5 on VM (Doc ID 2478446.1)” for a possible hit.
Run the following command to find the version of the kernel:
uname -a
The expected output is:
Linux victoria 4.1.12-94.6.4.el6uek.x86_64 #2 SMP Tue Aug 15 14:24:27 PDT 2017 x86_64 x86_64 x86_64 GNU/Linux
The kernel (4.1.12-94) is old enough to be affected by this bug.
Run the following command to see if the flag (scatter-gather) is not set:
ethtool -k eth1 | grep sca
The expected output is:
scatter-gather: on tx-scatter-gather: on tx-scatter-gather-fraglist: off [fixed]
No fix is applicable because the flag is set on.