VMware Cloud Community
Ilia_shapira
Contributor
Contributor

New Install ESXi 3.5 issue

I installed a few days ago the ESXi 3.5 on a server with Adaptec 3408 RAID

Everything was working fine, but the next morning from the install the server stoped responding, I was able to ping it but the local consloe was not responding and I was also not able to connect remotly. I restrted the server and again all day everything was working fine. Next morning again tthe same story.

I'm new to VMWARE so can anyone please tell me what can be the problem, where should I start looking and what should I do ?

Thanks.

Reply
0 Kudos
52 Replies
Deryni
Enthusiast
Enthusiast

I'm in the same boat. Trying to evaluate ESXi, but the management interfaces keep failing after a few hours (I originally thought it was nightly, but when I enabled alarms from Virtual Center, I noticed that it's failing earlier than that). I filed a service request (#1115504981) nearly three weeks ago and have had no response at all.

Ilia_shapira
Contributor
Contributor

Yep the support is really ..... I filed a service request more then two weeks ago and still got no word from them Smiley Sad

I still can't see any logic in the problems I'm having because sometimes the managment interface goes down after less then a day and sometimes it works for several days.

Reply
0 Kudos
Deryni
Enthusiast
Enthusiast

Mine seems to die within a few hours. Today's lasted 2.5 hours. It has yet to survive a night.

Reply
0 Kudos
cdickerson
Enthusiast
Enthusiast

I think I might of figured out what was going on in my circumstance. In the non ESXi world, it's best practice to create a separate VMKernel interface for vmotion. It appears in ESXi it wants you to use the Management Network for Vmotion. If you go into Vswitch0 under the Summary description, it will say Vmotion and IP Storage. So I enabled Vmotion on my Management Network, rebooted and have been solid since. I could hang my management console pretty much at will if I had a seperate vmkernel just for vmotion. I'm still in the testing phase to confirm this helped me, but it has been good for a couple of days and I can vmotion whenever and it doesn't hang.

Reply
0 Kudos
Ilia_shapira
Contributor
Contributor

I don't use Vmotion at all so I don't think this is a problem in my case

Reply
0 Kudos
Deryni
Enthusiast
Enthusiast

Same here. No VMotion, no HA, no extras. This was even happening to the ESXi server even when there was no Virtual Center in the picture. So it was being used as a glorified VMware Server, and the management interfaces were hanging.

Reply
0 Kudos
Deryni
Enthusiast
Enthusiast

Now interestingly, I reinstalled the machine with ESX server (not ESXi) and last night my management connections didn't die.

Reply
0 Kudos
Ilia_shapira
Contributor
Contributor

Of course its working, this is a totally different product.

It sounds that ESX and ESXi is something similar but in fact they are very diferent.

I'm sure that my problems will also be solved by installing ESX, but I'm looking for a low price solution and ESX costs much more plus you must buy support with it

Reply
0 Kudos
ocremel
Hot Shot
Hot Shot

Ilia, can you tell me a bit more about your machine ? How much memory does it have ? Did you start with blank disks (i.e. did ESXi auto-format them) ? Does the VI client health status page show any sensors ? few ? a lot ?

Reply
0 Kudos
Ilia_shapira
Contributor
Contributor

Its TYAN server.

It has two AMD Opteron CPUs, 8GB RAM and Adaptec RAID card.

Adaptec card is supported, but TYAN server is not fully suported so I dont see health satus.

Yes I started with blank disks.

Reply
0 Kudos
ocremel
Hot Shot
Hot Shot

I have a suspicion it is health monitoring that is causing problems. I will have someone get in touch with you on Monday for some diagnosis.

Reply
0 Kudos
Guillir
Enthusiast
Enthusiast

I had same issue on a compatible server on a customer last weekend

Dell PE 2950 III

2 x Quad Core 3.13GHz

8GB RAM

6 x 300GB 15kRPM

4 ethernet ports

ESXi build 94430 installable editon (installed on RAID 1)

Standalone ESXi with static IP and management interface on Vswitch0 with 3 ethernet ports in NIC Teaming

I could ping server, but console didn't respond as well "web interface".

VMs run OK during that time. I could open RDP connection on them and so on.

VI Client failed with "Server took to long to respond"

Telnet on port 902, 443 and 80 opened connection

After reboot ESXi, vnics were not bound to any vSwitch (actually were bound to an old vswitch name)

Guilherme Schäffer

Message was edited by: Guillir

Reply
0 Kudos
AKlotz
Contributor
Contributor

hello Guillir,

what storage controller are you using?

AL

Reply
0 Kudos
AKlotz
Contributor
Contributor

hello Guillir

what storge controller do you use?

AL

Reply
0 Kudos
Guillir
Enthusiast
Enthusiast

Hi Aklotz

The server has a Dell PERC6/i.

Raid 1 with disks 0 and 1

Raid 5 with disks 2, 3, 4 and 5

Reply
0 Kudos
AKlotz
Contributor
Contributor

hi guillir,

hm, i thought it could be something with the adaptec driver ... but than it does not look like this ...

thanks for your answer

Al

Reply
0 Kudos
skippermdj
Contributor
Contributor

I am having the same problem. Connectivity is lost at random intervals, usually several hours, sometimes after a couple of days. The virtual machines stay up and running, but the connection with the Virtual Infrastructure Client dies. I am running "headless", so I can't use the local console to restart the management network. I can ping the management network address, so I know the connection is up.

I am running VIC 2.5.0 build 84767 and ES3i build 82644. Hardware is a Supermicro X7DAE, Intel Quad-core 5410, 8gb ram, 6 intel e1000 ports bonded to two ethernet addresses: 3 NICS bonded to the VM Network and Management Network on vSwitch0, 2 NICS bonded to the Virtual Machine Network on vSwitch1, one unused NIC. One SATA drive as local boot device and iSCSI as datastore for the Virtual Machines. ATI Rage video card.

Currently on the 60-day evaluation. I'm surprised that, with several folks experiencing this problem, VMWare doesn't appear to have a patch or workaround. Any help will be greatly appreciated.

Thanks!

- Mark

Reply
0 Kudos
Ilia_shapira
Contributor
Contributor

Yes unfortunately the level of support is very very low.

I reported this problem here at forum a long time ago plus filed a service request, but still got no answer and no solution to the problem.

Reply
0 Kudos
xcheney
Contributor
Contributor

I am having the same problem. So happy to find you guys here.

My situation:

Supermicro Server with Adaptec RAID Card(RAID 5)

Installed ESXi (still in Evaluation)

VI Client can't connect to Host. Error:"took too long to response".

If I keep VIClient connected to HOST without closing it, it was OK. However, once I close VIClient, it is easy that can't reconnect to Host again, UNTIL restarting the HOST.

Same thing, when the Host is not responding VIClient, all the guests systems are running fine. I was thinking maybe it is a bug in ESXi that can't release/renew/reissue connections with VIClient.

I tried using VI Control Center to connect HOST, same thing as VIClient.

Since I am running the HOST as a production Server NOW, my hands are shaking....

If there is any one got solution, plz help....

Reply
0 Kudos
xcheney
Contributor
Contributor

No news??

Reply
0 Kudos