v1.4.8rcX Bug reports and comments

[gunther_01]

Does anyone have Multicast working with this version?
I have snooping enabled, and I'm still seeing multicast go out ports that don't even have a device connected that is requesting the stream.
It doesn't seem to be working correctly.

Of course I can disable MC on a port and it stops working (tried and true a couple times by loosing OSPF in the process as well). That's not a great solution in a lot of ways LOL

[jportman]

WS-8-150-DC with 1.4.7 - I was having trouble on reboot SFP disappearing. wispswitch-1.4.8rc7 seems to have made that stop, though I saw nothing in the release notes that would indicate that was targeted...... thanks?

Unit now has multiple weeks of uptime since the downgrade to 1.4.7

Currently up 21+ days.

[sirhc]

WS-8-150-DC with 1.4.7 - I was having trouble on reboot SFP disappearing. wispswitch-1.4.8rc7 seems to have made that stop, though I saw nothing in the release notes that would indicate that was targeted...... thanks?

Unit now has multiple weeks of uptime since the downgrade to 1.4.7

Currently up 21+ days.

Are you going to go on site and try v1.4.8rc7 ?

[tonym]

Testing out 1.4.8 rc10

CDP and LLDP seem to be working, but it isn't supplying any remote port data.

CDP is missing 'cdpCacheDevicePort'
LLDP is missing 'lldpRemPortId', 'lldpRemPortIdSubtype', and 'lldpRemPortDesc'

We use this information to draw links between ports on two devices.

[jarred.thornton]

Sorry my first comment is a small complaint.

Any chance we can get a lock button for the drag and drop feature. As cool as I thought when I read it. I have setup 5 switches(12 port and 24 port) so far since the update and cycling through U -> Q -> D -> T -> E on every port I find even being very careful the vlans tend to move. Even being able to lock the top line since its the management vlan and the importance it holds would be a great help. I just don't want to lose access to my poe switch on a live site.

Thanks considering this.
Jarred

[Eric Stern]

Sorry my first comment is a small complaint.

Any chance we can get a lock button for the drag and drop feature. As cool as I thought when I read it. I have setup 5 switches(12 port and 24 port) so far since the update and cycling through U -> Q -> D -> T -> E on every port I find even being very careful the vlans tend to move. Even being able to lock the top line since its the management vlan and the importance it holds would be a great help. I just don't want to lose access to my poe switch on a live site.

Thanks considering this.
Jarred

I'll fix this in rc12

[mike99]

Edit: P.S. 3 images attached even if it's hard to see at first.

MSTP is still totally broken in RC12. Noting in the instance seem to work.


MSTP instance per port priority just do nothing. The instance root bridge path will be the same as CIST path no matter the value use.
MSTP root bridge priority do only a cosmetic change. It will rewrite the "root id bridge" value on the switch you do the change but won't affect root switch path. On attached image "MSTP instance root bridge id cosmtetic.PNG", you can see with port is use to master path will the switch "root id bridge" is here own "bridge id". Exemple:

23# show stp
Bridge ID : 32768.EC-13-B2-61-B4-5C
Instance: CIST
Root Bridge ID: 32768.EC-13-B2-61-13-F6
Root Path Cost: 20000
Port Status Role
---- ----------- -----------
1 Forwarding Root
2 Discarding Alternate
3 Forwarding
4 Forwarding
5 Forwarding
6 Forwarding
7 Forwarding
8 Forwarding
9 Forwarding
10 Forwarding
11 Forwarding
12 Forwarding
13 Discarding
14 Discarding
Instance: 1
Root Bridge ID: 32769.EC-13-B2-61-B4-5C
Root Path Cost: 0
Port Status Role
---- ----------- -----------
1 Forwarding Master
2 Discarding Alternate

MSTP instance "root path cost" always show 0.

MSTP often dispaly their instance bridge ID in the field "root bridge ID" and in the "bridge ID field, display CIST "bridge ID" instead of the instance "bridge ID". It's like the value is writed in the wrong field.

[mike99]

I forgeted, no matter the version of STP, on down or disabled port, STP port state will initially display Forwarding insted of Discarding, or Disabled in STP version, until first state change. It's like it write this value on state change and default is Forwarding for every port. A init script should write those value since STP and RSTP/MSTP won't show the same value for a down port.

[Dave]

Mike99..thanks for your comments, Eric will look into it soon.

[mike99]

Ok, I digged a little further and it seem to be a region ID problem. The switch beahvior like if region are not matching from one switch to the other.

I try with a HP procurve 2530-24G, a Positron Access Osiris 1024F and a Netonix WS-12-250AC.

If I set same name and revision on the 3 switch, HP and Osiris can see each other on instance and master priority work fine for those 2 while Netonix still see itself as master and port forward follow CIST. Master priority change on the Netonix won't affect the HP and Osiris. If I set different name the 3 switchs, the 3 swtichs behavior just like the WISP switch (see those self as root and follow CIST forwarding).
Tryed back with 3 netonix, make sure it the same name (test), same revision (10), same vlans list (50). CIST behavior fine. Instance see there self as master and follow CIST forwarding.