Problem with EWA0 Speed

[rzamriv]

Hello, I have a problem with the EWA0 Device, I have changed the Speed with the MC LANCP Command, from 100 mbps to 10 mbps, but when I reboot the Alpha Station doesn't save that configuration, I reviewed the specifications but still in 100 mbps, and again I changed the speed manually, Do you know how I can save this configuration?, I have attached the procedure that I used.
$ mc lancp
LANCP> show dev/param ewa0

Device Characteristics EWA0:
Value Characteristic
----- --------------
Normal Controller mode
External Internal loopback mode
CSMA/CD Communication medium
16 Minimum receive buffers
32 Maximum receive buffers
No Full duplex enable
No Full duplex operational
TwistedPair Line media
100 Line speed (megabits/second)
LANCP> set device ewa0/speed=10
LANCP> show dev/param ewa0

Device Characteristics EWA0:
Value Characteristic
----- --------------
Normal Controller mode
External Internal loopback mode
CSMA/CD Communication medium
16 Minimum receive buffers
32 Maximum receive buffers
No Full duplex enable
No Full duplex operational
TwistedPair Line media
10 Line speed (megabits/second)
Thank you

[Steven Schweda]

> [...] the Alpha Station [...]

A model name/number might convey some useful
information.

> LANCP> show dev/param ewa0

Try "list" instead of "show", "define"
instead of "set"?

LANCP> help

[RBrown_1]

I suspect that SET only modifies the volatile database, and that DEFINE modifies the permanent database. Try using both DEFINE and SET to change both the permanent and volatile databases.

[Hoff]

A common approach is via the SRM console environment variables.

With LANCP, you need to specify both SET DEVICE and DEFINE DEVICE command to change the live configuration and for a change to "stick" over a reboot, respectively.

LANCP works like DECnet NCP, and it implicitly assumes the user knows the differences between the "volatile database" and the "permanent database"; this distinction was present with DECnet Phase IV, too. If you don't spot this reference and don't recognize it (and it's very easy to either not know about this stuff as DECnet is pretty old, or to simply not spot this reference in the help text) then the LANCP HELP text for the SET command doesn't really help you all that much.

[Jess Goodman]

Hoff wrote:
"A common approach is via the SRM console environment variables."

That's probably the easiest approach for DE500 devices. IIRC the console command would be:
>>>SET EWA_MODE "TwistedPair"
to go back to 100Mb you use
>>>SET EWA_MODE "FastFD".

However these commands do not work for Gigabit Controllers (DEGPA, DEGXA)

[Robert Brooks_1]

Why are you trying to do this?

Unless something is a bit strange with your configuration, it's usually best to set both
the ethernet adapter and the switch port to autonegotiate. (This advice comes from the guy who wrote ethernet drivers for VMS for a couple of decades [who is not me, but I've heard him say this repeatedly]).

If you're on a new-ish version of VMS, the $ SHOW DEVICE/FULL might be of interest.

-- Rob

[John Gillings]

Reinforcing Rob's comment.

You need to have a REALLY, REALLY good reason to set any ethernet port to anything other than AUTONEGOTIATE. These days that's pretty much true of all systems.

Note that any claims about OpenVMS problems with autonegotiate are NOT a good reason. For any hardware/software younger than 20 years it all "just works".

Hard set port speeds just create multiple potential misconfigurations, which can be difficult to detect.


[with hindsight, if I were redoing the autonegotiation algorithms I'd make it so if an autonegotiation request went unanswered, the port speed would be deliberately set, so a mismatched pair of ports is guaranteed to NOT WORK!]

[[on the other hand, there's a lot of people out there, including myself, who earn a significant fraction of their income from diagnosing and fixing duplex mismatched network connections!]]

[Volker Halle]

To add to John's and Rob's reply:

OpenVMS LAN engineering has even added error messages and error log entries, if the LAN driver detects/suspects a possible duplex and speed mismatch setting. This has been added in LAN patches for V7.3-2.

You can easily view these LAN driver messages on the console and with

$ MC LANCP SHOW DEVICE/INTERNAL_COUNTERS

at the bottom of the display for each LAN device.

%EWA0 possible duplex mode mismatch condition detected

Also see the following ITRC thread:

http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1108683

Volker.

[Richard W Hunt]

Oddly enough, I've run into this myself. We have a U.S. Government site so the "public bid" laws mean we have a mish-mash of different vendors for routers, firewalls, and network hubs. Even though I had the LANCP settings where I thought I wanted them, they didn't always take. That is when I had to reset the parameters associated with the SRM console settings. But then, it finally hit me.

Once you start the drivers in the console, some of that autonegotiation takes place right away, even before you reboot. And once you negotiate, it isn't so easy to force the card to re-negotiate. For some strange reason, our SRM console always wanted to negotiate HDX even though we knew better. Once I forced FastFD on each EWxx card from SRM, our problems went away.

Bottom line, set what you want in the SRM environment variables AND save the configuration for the cards in your OpenVMS database too. But then, I'm a "belt and suspenders" man myself, so I'm sometimes known for a slight overkill.

[Andy Bustamante]

rzamriv,

regarding ..the Alphastation..

What model Alpha system and what network interface? You can use

$ anal/sys
>clue config

to display a list of interfaces.

Second, what version of OpenVMS. Are you current with ECOs/patches? There used to be an issue with the LANCP database and console setting conflicting.