cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1186
Views
0
Helpful
3
Replies

Losing IPV4 admin connectivity on SG300-28P

Jeff Sterck
Level 1
Level 1

I have upgraded to FW 1.3.5.58 about 2 months ago.  Everything seems to be working as expected.  Just recently, I noticed that If I rebooted one of my IP phones that is powerd from the switch, no DHCP lease.  I currently have this connected to a Cisco 891 router which is doing the DHCP server function.  My voice vlan (200) and client vlan (100) both get DHCP this way.  I have had no issues with DHCP on the vlan 100, and was just noticing the issue with vlan 200.

So, I go to manage the switch, and I get no connectivity.  I have a management vlan (1) which is how I usually manage the switch, but cant ping the address.  the SG300 is doing L3 routing so I have an address on Vlan 100 as well, cannot ping that either.  Other than this, my network has been working like normal.  The odd thing is that I have vlan 1 and vlan 100 dual stacked with IPV6.  I can connect to each of the vlan 1 and vlan 100 IPV6 addresses.  When I log into the switch via IPV6, everything looks normal, no indication of issues in the logs.

Getting frustrated, I end up rebooting the switch.  LIke magic, I can ping V4 vlan 1 and V4 vlan 100, everything looks good.  DHCP passing to the 891 for vlan 200, phones are getting addresses.

24 hours later, I go to check the vlan 1 V4 connectivity, and nothing.  Vlan 100, nothing.  Reboot the switch again, works great.

Im noticing a lot of threads about ARP and seems like I could be running into this issue as well.  I recently downgrades to 1.3.0.62.  Ill watch that for a day or 2 and see how it does.

3 Replies 3

DJX995
Level 3
Level 3

I ran into this same exact issue on both my SG500-52P & my SG300-10.

After upgrading to 1.3.5.58...

My solution was to factory default the units (actually did it twice, just to make sure) and reconfigure from scratch: line by line.

I used the CLI to reconfigure.

My network has been very stable ever since I did that.

It's a big pain, but it's done now and it works.

Thanks for the heads up.  I ended up rolling back to the previous code (1.3.0.62) and so far so good.  When I get more time, Ill try your suggestion!

On all of our SG300-28P switches are doing this, the management interface is hanging sometime between a few hours to a few days after reboot.  When this occurs the switch continues to function, but the management IP does not respond to pings, http or ssh connections.  If I connect via serial cable, the switch is unable to ping out with an error "Ping: unable-to-send".

 

I'm running the current firmware version 1.3.7.18.  Is this a known bug?  Are there any specific configs that trigger this bug?  Are there any specific configs that might trigger this bug?  An unreachable sntp or syslog server?

 

Comparing sh tech between dead and working mgt interfaces, I only see this difference

------------------ buffers info ------------------
Name              Addr      Bsize    Bufn    Free Type
--------------------------------------------------------------------
SNLN           1FD1EBC     2560     344     306    Net Rx Long   <== Good

SNLN           1FD1EBC     2560     344       0    Net Rx Long  <== Bad
 

 

Any idea what the SNLN buffer is?  It is running out of memory the issue?