Fibre Channel (SAN)

Reply
Contributor
Posts: 39
Registered: ‎02-14-2006

Fabric Watch Best practice Values

Hi,

Frequent Contributor
Posts: 89
Registered: ‎05-12-2004

Re: Fabric Watch Best practice Values

I think that the best values to set for there areas are ones that will have meaning in your environment. On an ISL, typically one would want to know when rx/tx is getting to the point where additional ISL may be needed to handle the load. Of course one could also look at that time what is causing all the IO and re-adjust to be more localized, but that is a whole nother story. So, you really want to utilize your current ISL since adding a new one eats up ports. I think that a good number to set the rx/tx above alert to is 90% of the link bandwidth with a 10% buffer (traffic is typically bursty, so we don't want a ton of alerts when it is right around 90%, but if it drops below 80% and then jumps again above 90% that might be something to look at). Of course you may want to cosider different numbers depending on what your tolerances are, and what you are exactly looking to do with the alerts. I wouldn't bother with the below, in-between, or changed alerts for rx/tx since they really don't give alerts you of any issue that that you may have to deal with (unless you are concerned when there is little to no traffic on an ISL, but I would guess that at some time there would be "quiet" times on your SAN).CRC errors are also something that needs to be addressed based on the environment. If you are only seeing 100 frames per minute, 5 CRC errors would be 5% of the frames and maybe something that you would want to know about. On the other hand if you are seeing 100M frames then 5 CRC errors is a very small percentage. CRC error do occur, but they are rare, so you would need to set the number to be some percentage of your expected frames where there is some room for some errors, but not too much that you may miss a potentially failing SFP/cable.

Contributor
Posts: 39
Registered: ‎02-14-2006

Re: Fabric Watch Best practice Values

Many thanks for this detailed information.Chris

Occasional Contributor
Posts: 7
Registered: ‎07-06-2006

Re: Fabric Watch Best practice Values

I know this thread is kinda old but I wanted to point to a best pratice guide that is for the 12000 & 24000 Directors. Strange that this is not part of the Fabric Watch Adminstrator Guide.

N/A
Posts: 1
Registered: ‎12-07-2006

Re: Fabric Watch Best practice Values

I am trying to setup Fabric Watch to monitor mainly the hardware. I wanted the switch to send me email when there is hardware fault. I figured out I can configure most of the stuff using fwConfigure and fwMailCfg. But test email just never get through. I think that is because I did not have SMTP server defined on it. Question is where and how do I define the SMTP server?

Frequent Contributor
Posts: 78
Registered: ‎12-23-2005

Re: Fabric Watch Best practice Values

Hi Horpek,bernd

Contributor
Posts: 67
Registered: ‎08-29-2005

Re: Fabric Watch Best practice Values

Fabric Watch is actually seriously broken in this regard. Without a default mail server, it's pretty tough to get the mail out properly. .../Ed

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.