01-16-2018 08:04 AM
Hello,
I seem to have an issue where bandwidth between master <-> media server is only about 1Mbps.
Media server is 10/10 Mbit
Master server is 30/25 Mbit
Media server WAN optimization measured on both systems:
MediaDR.WANOptimization> Traffic 10 20
-------------------------------------- --------------------------------------
| OPTIMIZED | | NON-OPTIMIZED |
-------------------------------------- --------------------------------------
Time Offset MB/sec Mb/sec Kb/sec MB/sec Mb/sec Kb/sec
(sec) (avg) (avg) (avg) (avg) (avg) (avg)
----------- ---------- ---------- ---------- ---------- ---------- ----------
10 0.00 0 0 0.24 1 1931
20 0.00 0 0 0.24 1 1921
30 0.00 0 1 0.24 1 1907
40 0.00 0 1 0.24 1 1909
50 0.00 0 1 0.24 1 1947
60 0.00 0 2 0.24 1 1950
70 0.00 0 2 0.24 1 1914
80 0.00 0 0 0.25 1 1969
90 0.00 0 1 0.24 1 1940
100 0.00 0 1 0.29 2 2340
110 0.00 0 2 0.24 1 1923
120 0.00 0 0 0.24 1 1938
130 0.00 0 0 0.24 1 1928
140 0.00 0 1 0.24 1 1917
150 0.00 0 14 0.24 1 1932
160 0.00 0 2 0.24 1 1937
170 0.00 0 1 0.24 1 1944
180 0.00 0 1 0.24 1 1951
190 0.00 0 0 0.24 1 1946
200 0.00 0 0 0.24 1 1902
Master appliance:
Main.WANOptimization> Traffic 10 10
-------------------------------------- --------------------------------------
| OPTIMIZED | | NON-OPTIMIZED |
-------------------------------------- --------------------------------------
Time Offset MB/sec Mb/sec Kb/sec MB/sec Mb/sec Kb/sec
(sec) (avg) (avg) (avg) (avg) (avg) (avg)
----------- ---------- ---------- ---------- ---------- ---------- ----------
10 0.22 1 1751 0.07 0 552
20 0.22 1 1767 0.07 0 587
30 0.22 1 1759 0.07 0 542
40 0.22 1 1736 0.07 0 528
50 0.22 1 1743 0.08 0 604
60 0.22 1 1749 0.06 0 477
70 0.22 1 1729 0.07 0 552
WAN Optimisation is enabled on both appliances with the below parameters:
master:
32 15000 750 1 2 100000 125000 4 2 0 0 32768 0 1 0 4096 1 1 5000 2.1.1 13698 226 7921574 0 790656829 1212969133 1640042869408 861130574 1161517458629 351838559 478525410779 5 1 32 0 0 2000 25 1
Network Optimization Report
---------------------------
TCP connections optimized = 225
TCP connections not optimized (or not appropriate for optimization) = 0
Network Optimization = ENABLED
BONDING
-------
bond0 : 10Gb/s FC PLUGGED|UP 172.24.1.55 -- None --
Mode (802.3ad) Slaves ( eth6 eth7 )
media:
32 15000 750 1 2 100000 125000 4 2 0 0 32768 0 1 0 4096 1 1 5000 2.1.1 13698 89 1095138 0 901119849 40840673 22459245756 36617134 21437318279 4223539 1021927477 5 1 32 0 0 2000 25 1
Network Optimization Report
---------------------------
TCP connections optimized = 89
TCP connections not optimized (or not appropriate for optimization) = 0
Network Optimization = ENABLED
MediaDR.Network> LinkAggregation Status
BONDING
-------
bond0 : 1Gb/s TP PLUGGED|UP 172.24.1.77 -- None --
Mode (802.3ad) Slaves ( eth0 eth1 eth2 eth3 eth4 eth5 )
Also I have below shown in dmesg on media server, what could possibly be wrong etc ?
Could anybody please advise ?
bonding: bond0: An illegal loopback occurred on adapter (eth3). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
bonding: bond0: An illegal loopback occurred on adapter (eth5). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
bonding: bond0: An illegal loopback occurred on adapter (eth2). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
bonding: bond0: An illegal loopback occurred on adapter (eth0). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
bonding: bond0: An illegal loopback occurred on adapter (eth4). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
bonding: bond0: An illegal loopback occurred on adapter (eth1). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
01-16-2018 10:20 PM
Hi Unkn0wnn,
I haven't used WAN optimisation on my appliances, so cannot comment directly on that but do have a couple of observations from the info you've provided.
I am guessing that you have already asked your network team to review the switches, their config (including aggregation) and routing - based on the "illegal loopback" messages you are seeing?
The other observation is you haven't stated what model of appliance, or their software versions, but your conversation is tagged as version 7.5. If you are running that version then your appliances would be running version 2.5. Apart from that version being end of life and no longer supported, the use of Eth0 for anything other than initial config of the device was unsupported until version 2.6.0.2. I'm not sure if that is a part of your issue. Here are a couple of dicussions I've been involved in regarding this:
Hopefully someone else will be along with more info for you soon.
Hope this helps,
Steve
01-17-2018 03:51 AM
bonding: bond0: An illegal loopback occurred on adapter (eth3). Check the configuration to verify that all Adapters are connected to 802.3ad compliant switch ports
This is a OS message not related to Netbackup
802.3ad (VLAN taggning) is a protocol where multiple VLANS can be trunked to one NIC. Switch side need to be configured for VLAN tagging.
Red Hat has a technical article about the error message (required a login)
https://access.redhat.com/solutions/56604
Else just google for the above error messages.
01-17-2018 08:40 AM - edited 01-17-2018 09:11 AM
Hello, thank you for your help so far
Appliance Model is NetBackup Appliance 5220.
Appliance Version is 2.5.2.
I have asked my technical team to have a look at bandwidth which oscilates in ~1Mbps and config including routing.
From what I understand we must remove usage of interface eth0 from the aggregation (bond0) right ?
02-09-2018 06:10 AM
Still getting an illegal loopback error, cannot check with redhat as after registering they want me to buy subscription
02-11-2018 09:34 PM - edited 02-11-2018 09:35 PM
Hi Unkn0wnn,
Yeah, painful I know. When you do a basic Google search for that exact message, just about all the hits are from RedHat and you need the subscription. Almost looks like they are scrubbing any other answers from the Internet just so you buy a subscription
However, cutting some of the crap out of that message I came up with a hit from somewhere else that "appears" to have the RedHat answer in it
Have a look at the link below. You're looking for the 3rd reply (by Christopher Vanik). In between his two side-by-side screenshots of his console, and the network diagram there is a small hard to read graphic, which is directly under this text - "It seems like RHEL documented this issue previously, but I'm sure Cumulus VX switchports are 802.3ad compliant.". Have a look (double-click the graphic to enlarge) - I reckon that is the RedHat answer:
Hope it helps,
Steve
02-12-2018 10:04 AM
Thank you for that. I will check configuration of the switches on both sides and then will let you know