• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Switch Troubles.

M

ManateeMatt

Limp Gawd
Joined
May 27, 2009
Messages
145
I have a Netgear gsm7248 switch that's giving me a lot of headaches when trying to pass ISCSI traffic from my File server to my VM server. It only appears to be an issue with this switch, and only when I try to use ISCSI traffic.

The odd thing is, it only looks like it drops packets from Switch to VM server and not Switch to FS.

My *very* scientific approach to this was ping the server for along time before I started the VMware software ISCSI connector. Pings would come back in less then 1ms, but as soon as I told it to start using ISCSI ping would loose aprox 10 to 15% of packets

I have tried the following

Different cables
Different nics in both the VM server and FS
Different ports and port clusters
Resetting the config, and starting again
Trying new/old firmware for the router.
Tried with only VM server and FS connected.


Specs of VM box and FS

VM Server
Tyan S2912 with V4.0 bios
2x AMD 8356 @ 2.3GHZ
4x 4GB of some noname ramz
2 port Pro/1000GT
NORCO RPC-250
ESXi is running off of a 2GB flash drive.

File Server
Gigabyte p35-ds3l
Intel e2200
2x 2GB of noname ramz
Rocket Raid 2680
4x500GB Western Digital Black for VMware (raid 10)
3x 1.5TB Western Digital Green for Storage (raid 5)
Ubuntu 9.10
 
What I don't understand is how you concluded that the Netgear SW is the problem. Did you use another SW (that you may have forgotten to mention in your post?)

Do you have redundant connection for the VM box (meaning the box has 2 connections to the same SW, or same VLAN)?
One thing I would check is if there are any updates / known issues with VMWare's ISCSI controller implementation.
 
@ megabit

I forgot to post that it works on using my DIR-655 router/switch/AP thing. But thats a basic, un-managed switch.

No redundant connections at all right now, just 1 connection for ISCSI traffic, and 1 connecton for VMware VSwitch.

@ The Captain

Thanks for the suggestion. Will keep that in mind next time. Right now, not VLAN separation at all, but i will be doing that later.


Didn't want to start complicating things with Vlans and more complex configs right now. I just wanted to see if it was going to work.
 
I only had to set up iSCSI once when the technology was new but not only did we have separate VLANs, we had completely separate switches for the traffic.

It's pretty easy to choke out a low-end gig switch trying to do what is described above.

This sounds like a job for JUMBO FRAME!!!11 :)
 
well disabling STP look liked it helped for a bit, but the problem looks like it has returned.

ARG.

It only seems to happen when ksoftirqd_CPU0 gets stuck at anywhere between 50 and 90%

I will enable jumbo frames and see what happens, I think it might be time to find a newer/better switch.

I have been looking at the HP 1810g and the Linksys/Cisco SRW2024.

I am willing to go around 400 - 500 Canadian for a Gig switch, min of 24 ports.

I dont need uber enterprise class, but i want one that doesn't suck.
 
I'd probably go with the HP if I were you. I don't have any experience with them, and I am sick of hearing my friends rant and rave about how amazing they are.

I do have experience with the SGE2000 and SGE2000P Switches and the firmware was junk. Cisco released an update in December that allegedly fixes all the web interface bugs, but I already set it up through command line anyways, so why fix what ain't broke? Those switches have been running reliably for over a year with no packet loss or sudden failures. I probably just jinxed myself...

The SRW20xx series has multiple reports of major reliability problems, think twice before purchasing that one.
 
Last edited:
I'd go with the HP ProCurve 1810-24G. I have an 1800-24G and it works just fine for me. ProCurve has been growing on me the last six months or so as I've had to use them for a few smaller VMware setups where the clients just couldn't afford Cisco. I still wouldn't use one as the core for a large network but they work just fine for smaller setups.
 
I think that's what I am going to go with. I have had enough of this stupid switch to last a life time.

It just doesn't make any sense why this switch does not play nice with my setup, when a dlink DIR-655 works without any issues.

I guess its managed vs unmanned. Managed is trying to do to much I guess, and the netgear just does not play nice.
Posted via [H] Mobile Device
 
You must log in or register to reply here.
Back
Top