Network uplink redundancy lost vmware как исправить
I often get these alerts on one of our servers in our environment. We run VMware ESX 4.1 on 5x Dell PowerEdge Servers each of similar spec, they all have 2x NICs that plug into a Cisco Infrastructure (6509 Chassis). They are all configured the same so I don’t have reason to suspect a mis-configuration. This alert could be related to a hardware problem as I don’t see anything on the Cisco side stating that the interface has gone down.
My first question is where in the logs I can check for more information on this issue? I have checked the /var/log/vmkwarning log but there is nothing to indicate that the connection has gone down.
From the Client View I can see the following events:
Lost uplink redundancy on virtual switch «vSwitch0». Physical NIC vmnic1 is down. Affected portgroups:»Vlan1xx», «Vlan1xx», «Vlan1xx», «Service Console», «VMkernel». warning 26/05/2011 15:22:23 esxserver2.domain
Uplink redundancy restored on virtual switch «vSwitch0», portgroups: «Vlan1xx», «Vlan1xx», «Vlan1xx», «Service Console», «VMkernel». Physical NIC vmnic1 is up. info 26/05/2011 15:22:27 esxserver2.domain
This constant issue turned serious this morning when the VMNIC1 changed speed to 100MB, this caused the Cisco end to drop the interface and I strangely managed to lose network connectivity to half the Virtual Machines on the host? It seems the Cisco side is set to auto so I will probably change this but it has never been an issue prior to this morning.
Prior to logging a call with Dell I could do with gathering the evidence required to prove hardware failure or bottoming out what the issue may be.
I’ve had similar issues as far as the network speed dropping to 10 or 100. First thing is that I make sure both sides are set to Auto if it’s 1 GB links. Second, is that I have our hardware team reseat the cabling all the way back to the switch. 9/10 times that has resolved it. The root cause is usually the cabling was a little loose, not snapped all the way in, or the cable was bent forcing a pin or two to lose contact with the port.
In a few rare cases they were due to bad patch cables.
This may not be good for you since you only have 2 ports, but on our servers I’ve also modified the host profile to fail a port if the link speed is below 999MB, that way it forces all of the traffic over to the other port without dropping speed as the switches try to renegotiate.
But since Cisco it does NOT disable the NIC and instead only marks it «non-participating» on the etherchannel then the vswitch still attempts to route guest traffic through that 100 mbit NIC and those packets are dropped on the Cisco side because that port is now «non-participating».
The long-term resolution may be to reconfigure all those ports as individual trunked access ports instead of bonding them into an etherchannel; but in the meantime I am looking for a way to actively monitor for any changes in vmnic negotiated speed, and preferably an automated way to fail that port.
Network uplink redundancy lost vmware как исправить
i am receiving below mentioned error on one of my host and the «STATUS LED’s) in front is on. kindly suggest.
1) Network connectivity lost
2) network uplink redundancy lost
If your host is still connected to the vCenter inventory or you can browse into the ESXi Host client go to the Physical Adapters section and you most probably see that there one of your Physical NICs is down. And log into the Administrative Console of your server from whichever vendor is and look for some physical devices alarms.
Both are default alarms of vCenter for Host network connectivity. Most likely someone pull the cable or misconfigurated the phys. Switchport.
I have Hosts with several Dual Port SFP+ NICS and often when rebooting a Server because of Patching the ports doesnt become fast enough active again when ESXi reboots. So i got warnings like you that some nic port isnt connected but it is when checking the vNetwork on the host.
So as mentioned just check the network cards or your vswitch topology if all uplinks have links and when not which one is the missing one.
Более 5490 заметок о виртуализации, виртуальных машинах VMware, Microsoft и Xen, а также Kubernetes
VM Guru / Articles / Решение проблемы потери сетевого подключения у виртуальных машин VMware в облаке IaaS.
Решение проблемы потери сетевого подключения у виртуальных машин VMware в облаке IaaS.
Решение проблемы потери сетевого подключения у виртуальных машин VMware в облаке IaaS.
Автор: ИТ-ГРАД Дата: 24/10/2016
Представляем очередной гостевой пост компании ИТ-ГРАД, посвященный инфраструктуре VMware vCloud Director. Посмотрите их интереснейший блог об облачном IaaS.
Как часто вы сталкиваетесь с проблемами в облачном окружении, устранить которые можно обходными путями по причине отсутствия конечного решения? Одна из таких проблем – потеря сетевого подключения у виртуальных машин на базе ОС Windows Server 2012/R2 с сетевыми адаптерами E1000/E1000e в облаке IaaS.
Признаки проблемы
Распознать проблему просто, ведь ВМ теряет подключение к сети, восстановить которое можно остановкой и последующим включением виртуальной машины. О сетевых неполадках также свидетельствуют записи в журналах Event Logs (System), которые содержат следующую информацию: «Intel(R) 82574L Gigabit Network Connection Network link is disconnected» или «Intel(R) 82574L Gigabit Network Connection Network link has been established at 1Gbps full duplex».
Эта проблема давно известна и даже описана в базе знаний VMware, но окончательного решения вендор пока не предложил. Вместо этого рекомендуется использовать временный вариант устранения неисправности, о котором подробнее расскажем ниже.
Вариант решения проблемы
Решить озвученную проблему предлагается путем замены адаптера E1000/E1000e на VMXNET3. Но прежде чем вносить изменения, убедитесь, что в виртуальной машине установлены инструменты VMware Tools. Дело в том, что драйвер для адаптера VMXNET3 поставляется только в составе с VMware Tools.
Убедиться, что VMware Tools установлены, можно во вкладке General свойств виртуальной машины консоли vCloud Director:
Свойства ВМ консоли vCloud Director
Зачем менять один адаптер на другой? Чтобы стало понятнее, предлагаем пройтись по характеристикам каждого из устройств.
Разница между адаптерами VMXNET3 и E1000/E1000e
При сравнении адаптеров VMXNET3 и E1000/E1000e мы имеем дело с паравиртуализированными и эмулируемыми сетевыми картами соответственно. Но в чем отличие?
VMXNET3 представляет собой паравиртуализированный сетевой адаптер с прицелом на максимальную производительность. И часть того, что ранее эмулировалось, теперь передается в физическое устройство напрямую. Драйверы VMXNET3 входят в состав VMware Tools и поддерживаются большинством гостевых операционных систем.
Сравнение пропускной способности адаптеров
Что касается производительности, E1000/E1000e значительно уступают VMXNET3. Кроме того, VMXNET3 отличается большей стабильностью и требует под себя меньше процессорных ресурсов.
Guest customization
Прежде чем начать замену адаптера, следует убедиться еще кое в чем. Например, проверить, что не активирован механизм Guest customization. Отключить эту опцию можно в свойствах виртуальной машины, перейдя в закладку Guest OS Customization консоли vCloud Director.
Обзор параметров закладки Guest OS Custamization
И еще несколько полезных рекомендаций, на которые стоит обратить внимание:
Замена адаптера E1000/E1000e на ВМ
А теперь перейдем к решению проблемы. Чтобы произвести замену адаптера на виртуальной машине, выполните следующие действия:
Добавление нового адаптера
Внимание! Если vCloud Director предлагает использовать Guest Customization, нажмите No.
В этом кейсе мы рассмотрели вариант решения проблемы, связанной с потерей сетевого подключения в виртуальных машинах на базе Windows Server 2012/R2 с сетевыми адаптерами E1000/E1000e в облаке IaaS. А с какими проблемами в виртуальном окружении сталкиваетесь вы?
Чтобы оставлять комментарии, вы должны быть зарегистрированы на сайте.
It’s amazing how much is going on when you dig through logs. On this occasion I was looking at  “tasks & events” of a host and noticed a lot of network errors.
Alarm вЂNetwork uplink redundancy lost’ on triggered an action
The error was occurring every 5 minutes. This was made visual with the use of Log Insight. My new favourite tool.
I couldn’t find anything wrong with this particular ESXi host, vSwitch or uplink. It had the same configuration as all the other hosts in the cluster.
The fix was to go to the top level where the alarm is defined, Edit Settings, disable the alarm, then go back and re-enable it.
After that, the errors stopped appearing.
Comments
Comment by Anonymous on 2014-02-24 17:45:42 +0000
I have same issue at the moment, errors also stopped for me by following your steps however on checking logs within host, issue is still there.
So you should remove this blog, which provides a wrong soln, which only stops events from coming.
Comment by daunce on 2014-02-24 17:53:57 +0000
Of course you need to re-enable the alarm.
In my case it was a false positive.
Your situation may be different. If it continues, contact VMware support.
Comment by Anonymous on 2015-03-27 12:13:48 +0000
Thanks! Turning this alarm off then back on fixed the same issue here. Wonder what might trigger that.
Comment by Anonymous on 2015-07-14 02:30:07 +0000
Ran the most recent updates for 5.5 and this alarm triggered afterwords. Disabling then re-enabling still fixed it! Thanks for posting the tip.
Comment by MIke on 2017-05-12 18:57:15 +0000
Issue still occurring in esxi 6.0 just disabled and then reenabled we will see what happens
Comment by Kamran Zafar Syed on 2018-03-05 13:59:05 +0000
thanks man that did the trick it is VC 5.0. And email keeps popping after every 5 minutes.
Comment by Tony on 2018-11-20 05:43:29 +0000
This worked great on an ESXi 5.1.0 host. We had a network switch problem that caused links to go up and down. The alarms continued after the switch stack restart even though the connections were all restored.
Network uplink redundancy lost vmware как исправить
I was getting the above errors every 2 mins, but I did not find any issues with my network.
I updated my ESXi to latest 5.5 with all updates with no help.
I have reffer to this simple solution of Alarms enabled/disabled, and I am no longer getting this error:
Can any one confirm me, if my issue is really resolved or do I still need to worry, is their way to double check?
This error generally gives only if your uplink goes down, it can either due to physical connection unplugged or packet loss in a certain amount of time. What is the redundant policy that you have set it. Can you try by explicit fail over. This would give a better picture if you really looses the network.
You should check for port speed/duplex mismatches on your vmnics and physical switch ports.
Make sure the vmnics and the physical switch ports are all set to auto negotiate
One of the Nic was on Auto and other not, I changed it to Auto as well, and now errors are back.
Also one of the ISCSI port nic teaming was different to other, I changed it to active-active as well.
If this is a production environment tread very carefully making changes to your storage network.
Looking at your config you should configure it like this providing you are happy the connectivity is set up correct:
ISCSI-1 vmnic2 active, vmnic3 unused
ISCSI-2 vmnic3 active, vmnic2 unused
Check out this short whitepaper on iscsi multipathing as it may help you in troubleshooting further
