Skip to main content
Start a Conversation

Unsolved

This post is more than 5 years old

N

nlspears

52 Posts

43233

September 6th, 2007 14:00

Link Up and Down

Hi was looking in my RAM Log on one of the 3448 switches and noticed several links going from up to down and back up status. An example is below:
 
%LINK-W-Down: E37
%LINK-I-Up: E37
%LINK-W-Down: E37
 
Is this normal?

Crzywolf

93 Posts

September 14th, 2007 16:00

What do you have connected to that port?  Is the speed and duplex hard set or set to auto-negotiate?

nonarKitten

14 Posts

September 18th, 2007 20:00

Having the same problem here.
 
I've been unable to find a commonality of those ports that fail. For example, some are just workstation 100b ports on our 3448's - there's what appears to be a random though continuous port interruption that lasts a couple of seconds up to a couple of minutes. Without logging we might not have even noticed.
 
On the 5324's (which are used as our backbone switches) its more painfully obvious, especially with the trunk ports. For example, we'll get the following sequence of errors:
 
  187   2147483175   21-Aug-2007 08:18:23    Informational  %LINK-I-Up:  ch5          
  188   2147483176   21-Aug-2007 08:18:23    Informational  %LINK-I-Up:  g23          
  189   2147483177   21-Aug-2007 08:18:23    Informational  %TRUNK-I-PORTADDED: Port g23 added to ch5          
  190   2147483178   21-Aug-2007 08:18:20    Warning  %LINK-W-Down:  ch5          
  191   2147483179   21-Aug-2007 08:18:20    Warning  %LINK-W-Down:  g23          
  192   2147483180   21-Aug-2007 08:18:20    Warning  %TRUNK-W-PORTDOWN: Port g23 removed from ch5: port is down/notPresent          
Of course, this happens on our iSCSI servers too, causing routine failures in the volumes. Windows is pretty tollerant and will happily wiat 60 seconds for the volume to come back online. They haven't a couple of times, forcing manual intervention and a thourough chkdsk.
 
I've checked all the port settings, they seem okay. It's odd that many are MDI and others are MDIX (thinking that the electrician got some reversed), but the pattern of this is inconsistent with the ports that fail. I've also tried disabling STP though it still appears to be on at the path lengths are getting updated (again, they vary, but not consistently with the interruptions).
 
All switches have auto-negotiation and auto-MDI/X enabled (thinking I should disable this too for the iSCSI), but have pretty much everything else turned off including storm control, back-pressure, flow-control, multicast filtering and snooping. We have no VLANs, GVRP is globally diabled (though Dynamic VLAN and GVRP registration are enabled on the ports - this is a confusing interface; I hope the global means "override"). The 5324's (the most critical) have the following rev's:
 
Software Version  1.0.0.45   
Boot Version  1.0.0.21   
Hardware Version  00.00.02 
For sake of explanation, here's our basic network topology
 
iSCSI Targets
 - 5324 (SAN)
ESX Servers
 - 5324 (LAN)
    - 3448's (three)
    - 2748
 
All very basic - so I'm not sure why we're having these connectivity problems. Any help would be appreciated. Thanks.
 
 
 
 

basheerpt

77 Posts

October 23rd, 2010 05:00

Hi

Thanks for your post, I am also experiencing the same issue with my 5424 switch, I have 2 switches, connected to Dell equallogic SAN . I lose connectivity often with one of the switch and all connectivity to san is lost. I see the errors in the log (one of the port up and down). I just wonder, if i agree this ports is faulty, it will stop the entire switche's connectivity?!

Lately, if you could find a solution, kindly share with me.

Thank you

BASHEER

bh1633

909 Posts

October 25th, 2010 08:00

Please post your configuration, description of your network, log file output. 

Links going up and down are often HW issues.  To debug, One at a time in this order:, change ports, change cables, change transceivers,change switch.

View More

View All

No Events found!

Top