| Author |
Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 17 2010 17:02 |
|
|
I am using a PersonalAlpha-created disk image which is using a network configuration that worked under PA. Using FreeAXP, there is no network device other than LO0 and I seem to recall that, somewhere, there was documentation of how to "convert" network settings as device names are different between the two emulators.
I am using a dedicated network device and can join a cluster successfully, but TCP/IP fails with the SE0: device and hence I have no TCP/IP.
Cheers |
|
|
| Author |
RE: configure we0 interface in TCPIP |
VolkerHalle
Member
Posts: 104
Location: Germany
Joined: 02.04.10 |
| Posted on May 17 2010 19:44 |
|
|
Gremlin,
FreeAXP emulates a DE500 LAN interface, which will show up as EWA0 under OpenVMS and as we0 in TCPIP.
Use @TCPIP$CONFIG -> 1 Core environment -> 2 Interfaces and configure we0
Volker. |
 |
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 17 2010 20:02 |
|
|
Hi
That made no difference, but I went into LANCP, deleted SE0, added WE0, rebooted and then TCPIP$CONFIG found WE0.
However, still not working, probably because of my configuration....I am running FreeAXP under Windows2008 under Fusion for OSX!!! OSX is on a Mac Mini, therefore only a single network port. While this works fine for PersonalAlpha with their NDIS driver, it works poorly for FreeAXP. So, I added a USB/CAT5 converter which the VMware Windows can see and use, removed all bindings from it and configured it as the NIC for FreeAXP.
However, no network communication occurs through the device with VMS. If I enable TCP/IP and MS protocols again, the Windows host has no problems - so, am I going one step too far? |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
iamcamiel
Moderator
Posts: 25
Location: Wageningen, The Netherlands
Joined: 06.04.08 |
| Posted on May 19 2010 04:06 |
|
|
Hello Gremlin,
Just for clarification: you can join a cluster, but no TCP/IP from the emulator running under Fusion? If so, that's an indication that networking to the outside world is working.
Camiel. |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 19 2010 16:48 |
|
|
Oops - double checked....the cluster join request is sent, but nothing happens. So, there is no network activity across the link by VMS.
Analysis indicates that Fusion sees the device (I can add it to a VM), the VM sees the device (I can give it an IP address within Windows and ping it, send/receive information using it and unbind all services/protocols from it), FreeAXP sees it (I can assign it as the network device for the emulator) and I think that OpenVMS sees it as I can see the device in SRM, change its mode to FastFD and within OpenVMS, I can do a show dev/fu EWA0 and it shows as a DE500.
Its just that no network traffic is moving across it!! If instead I use the MacMini inbuilt network card (bound to IP etc within Windows), I get network traffic in OpenVMS but it clashes with everything else - running dedicated is the way to go......if I can get it to work. Again, works seamlessly under PA. |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 19 2010 17:51 |
|
|
Additional info:
On PA, there are two network devices, SEA0 (Lance device) and EWA0 - on FreeAXP there is no SEA0 device as nothing is bound to the network card in the Windows host. Could this be part of the problem? Recall, there are two NICs available in the host and I am dedicating one to FreeAXP with nothing bound to it. |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
VolkerHalle
Member
Posts: 104
Location: Germany
Joined: 02.04.10 |
| Posted on May 19 2010 19:15 |
|
|
Gremlin,
I can assure you, that there is no EWA0 device on PersonalAlpha, only an ESA0 device. In TCPIP interface naming, you could only configure SE0 on PersonalAlpha and WE0 on FreeAXP.
You can easily confirm this with $ SHOW DEV EW and $ SHOW DEV ES
and $ TCPIP SHOW INT. If you use $ TCPIP SHOW CONF INT, you may well see both se0 and we0 in the permanent database on disk, but you'll get an error about a missing device, when starting TCPIP.
Please note that network card drivers (other than on OpenVMS) may get confused, when you're starting DECnet with Phase IV address, as the MAC address needs to be changed.
May I suggest, that you try to configure the TCPIP interface on both PersonalAlpha and FreeAXP to use DHCP for configuration of the IP address and parameters ? This is the most flexible solution and allows you to boot the same OpenVMS disk in both emulators. Alternatively, you could also define both interfaces with fixed addresses in the TCPIP database. You have to do these configurations individually running under each emulator. You can even configure different hostnames with HOSTNAME.we0 and HOSTNAME.se0, so that your system shows up as PersonalAlpha.your.domain or freeaxp.your.domain depending on which emulator you boot your OpenVMS system disk on.
Volker.
Edited by VolkerHalle on May 19 2010 19:57 |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 19 2010 23:34 |
|
|
Hi
They are already configured as fixed IP addresses and ESA0 is the correct device name - I am having finger troubles!!
DECnet may be the culprit - I don't remember anything in some documentation about it, now I suppose I need to find it. I remember the need to change the MAC address is there was more than one NIC configured for VMS, but I suspect there is only one EWA0. I am happy to boot this system under FreeAXP only, as long as I can sort out the network issue.
[edit] checked the documentation and it mentions MAC issues if there are multiple FreeAXP instances on a single network (not in this case), running on a wireless card (no) and running two NICs (no). Can't see how DECnet would be a problem on a single dedicated NIC with fixed IP addresses. Also, as cluster join messages aren't getting out, I guess the problem is not IP related - it must be much earlier and much deeper.
Edited by gremlin on May 20 2010 00:30 |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
Shane Holloway
Member
Posts: 2
Joined: 31.03.10 |
| Posted on May 20 2010 02:27 |
|
|
Gremlin,
I believe you have to have two network cards configured on your Fusion-based Windows system to use FreeAXP's networking driver to provide networking support to emulated environment. Under Parallels VM, I just add a second network adapter to my system configuration.
Hope that helps!
-Shane |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 20 2010 02:51 |
|
|
Hi Shane
Thanks for that - Fusion has two adapters configured - one is the in-built Gigabit network port, shared by the Windows host. The other is a USB-CAT5 adapter that fusion sees and is also seen by the Windows host, but is not bound to anything within Windows. It is this adapter that is selected for exclusive use by FreeAXP.
Volker,
$ tcpip sho conf int
Interface: LO0
IP_Addr: 127.0.0.1 NETWRK: 255.0.0.0 BRDCST:
Interface: SE0
IP_Addr: 192.168.100.202 NETWRK: 255.255.255.0 BRDCST: 192.168.100.255
Interface: WE0
IP_Addr: 192.168.100.202 NETWRK: 255.255.255.0 BRDCST: 192.168.100.255
The system can ping itself, nothing else. Hosts and gateway are configured, as are DNS and routing. Remember, this is a PA image in which networking was fine until I tried to run it under FreeAXP. MAC address of EWA0 is 08-00-2b-de-50-00. There is no other identical MAC address on this network segment.
Cheers
Edited by gremlin on May 20 2010 02:52 |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 20 2010 02:57 |
|
|
Hi all
DECnet is emerging as the culprit....
LANCP>
**** OpenVMS Alpha Operating System V8.3 - BUGCHECK ****
** Bugcheck code = 000007CC: NETDLLERR, DECnet Datalink Layer detected a fatal error
** Crash CPU: 00000000 Primary CPU: 00000000 Node Name: GAMMA
** Supported CPU count: 00000002
** Active CPUs: 00000000.00000001
** Current Process: NULL
** Current PSB ID: 00000001
** Image Name:
**** Starting compressed selective memory dump at 20-MAY-2010 22:52...
|
|
|
| Author |
RE: Network issues with PersonalAlpha container |
Bruce Claremont
Moderator
Posts: 623
Joined: 07.01.10 |
| Posted on May 20 2010 17:06 |
|
|
|
You are running FreeAXP 271, correct? |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
VolkerHalle
Member
Posts: 104
Location: Germany
Joined: 02.04.10 |
| Posted on May 20 2010 19:18 |
|
|
Gremlin,
when starting DECnet Phase IV or DECnet-OSI with Phase IV compatible adressing, the MAC address of the LAN interface is being changed to the DECnet MAC address (AA-00-04-00-xx-yy). You need to start DECnet before starting TCPIP. Changing of the MAC address may 'confuse' the underlying network stack(s). You can check the current MAC address with SDA> SHOW LAN/FULL/DEV=EWA (look for Physical address at the bottom of the first page shown).
That the system can ping itself, does not prove anything. Such an ICMP packet never reaches the LAN interfaces.
You cannot draw any conclusions from just a NETDLLERR crash, especially if you don't provide more information from the CLUE file (especially module and offset of the crash). Feel free to send me the full CLUE file (CLUE$COLLECT:CLUE$node_ddmmyy_hhmm.LIS).
Volker. |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 20 2010 19:18 |
|
|
Yes, downloaded it about four days ago. It automatically removed the earlier version, or so it said  |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 20 2010 20:53 |
|
|
Hi Volker
Just scratching my head about how to get the file to you - there being no network connection the the machine.....and it is too big to capture it in screen buffer I think.... |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 20 2010 21:01 |
|
|
LAN Data Structures
-------------------
-- EWA Device Information 20-MAY-2010 23:50:59 --
LSB address 8385C000 Driver code address 81DFBD80
Driver version 00000001.08310124 Device1 code address 00000000
Device1 version 00000000.00000000 Device2 code address 00000000
Device2 version 00000000.00000000 LAN code address 81DFD400
LAN version 00000001.08310190 DLL type CSMACD
Device name EW_DE500 MOP name EWA
MOP ID 203 HW serial Not supplied
HW version 02000041 Promiscuous mode OFF
Controller mode NORMAL Promiscuous UCB 00000000
Internal loopback OFF All multicast state OFF
Hardware address 08-00-2B-DE-50-00 CRC generation mode ON
Physical address 08-00-2B-DE-50-00 Full Duplex Enable ON
Active unit count 7 Full Duplex State ON
Line speed 100
Flags: 00000140 TIMER,FAILDEV
Char: 000000C4 MULSEG,LINKVALID,AUTOVALID
Press RETURN for more.
|
|
|
| Author |
RE: Network issues with PersonalAlpha container |
VolkerHalle
Member
Posts: 104
Location: Germany
Joined: 02.04.10 |
| Posted on May 20 2010 21:25 |
|
|
Gremlin,
re: SDA> SHOW LAN output
you've either NOT started DECnet or you're not running DECnet Phase IV compatible protocol. SDA> SHOW LAN would show all protocols running on your LAN adapter. Which version of DECnet are you running:
$ MC NCP SHOW EXE
or
$ MC NCL SHOW IMPLEMENTATION
re: CLUE file
Look at my profile for my mail address. The CLUE file is only a couple of pages of plain ASCII text. Use cut & paste or the Putty logging feature and TYPE/NOPAGE the CLUE file.
Volker.
Edited by VolkerHalle on May 20 2010 22:29 |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 21 2010 01:15 |
|
|
Show LAN
LAN Data Structures
-------------------
-- EWA Device Summary 20-MAY-2010 23:50:59 --
LSB address = 8385C000
Device state = 00000103 RUN,INITED,RING_AVAIL
UCB UCB Addr Fmt Value Client State
--- -------- --- ----- ------ -----------
EWA0 82147F40
EWA2 8214C780 Eth 60-07 SCA 0017 STRTN,LEN,UNIQ,STRTD
EWA5 821E6AC0 802E 08-00-2B-80-3C DNAME 0017 STRTN,LEN,UNIQ,STRTD
EWA6 8221D2C0
LAN Data Structures
-------------------
-- EWA Device Summary 20-MAY-2010 23:50:59 --
LSB address = 8385C000
Device state = 00000103 RUN,INITED,RING_AVAIL
UCB UCB Addr Fmt Value Client State
--- -------- --- ----- ------ -----------
EWA0 82147F40
EWA2 8214C780 Eth 60-07 SCA 0017 STRTN,LEN,UNIQ,STRTD
EWA5 821E6AC0 802E 08-00-2B-80-3C DNAME 0017 STRTN,LEN,UNIQ,STRTD
EWA6 8221D2C0 Eth 80-3C DNAME 0017 STRTN,LEN,UNIQ,STRTD
EWA7 82254400 802E 08-00-2B-80-3E DTIME 0017 STRTN,LEN,UNIQ,STRTD
EWA9 8225EAC0 Eth 08-00 IP 0015 STRTN,UNIQ,STRTD
EWA10 822B3F80 Eth 08-06 ARP 0015 STRTN,UNIQ,STRTD
EWA11 822B4400 Eth 86-DD IPV6 0015 STRTN,UNIQ,STRTD
EWA12 822CC280 Eth 00-00 0000
EWA13 822D4C00 Eth 00-00 0000
EWA14 822E3040 Eth 00-00 0000
SDA>
GAMMA::[SYSERR] > mc ncp sho exe
Node Volatile Summary as of 21-MAY-2010 21:12:47
Executor node = 1.3 (GAMMA)
State = on
Identification = DECnet-OSI for OpenVMS
Clue coming soon |
|
|
| Author |
RE: NOT running DECnet routing yet.. |
VolkerHalle
Member
Posts: 104
Location: Germany
Joined: 02.04.10 |
| Posted on May 21 2010 02:13 |
|
|
Gremlin,
you haven't even started the DECnet routing protocol on this LAN adapter yet. Assuming it's still defined to use the ESA0 LAN interface (from PersonalAlpha).
Check SYS$MANAGER:NET$CMSACD_STARTUP.NCL:
...
CREATE NODE 0 CSMA-CD STATION CSMACD-0 COMMUNICATION PORT = ESA
...
Just change ESA to EWA and reboot or use NCL to change the information in the running system.
Also check NET$ROUTING_STARTUP.NCL for
SET NODE 0 ROUTING CIRCUIT CSMACD-0 ENABLE PHASEIV ADDRESS = TRUE
This will force the MAC address of the LAN adapter to be changed. If you don't like this, make it: PHASEIV ADDRESS = FALSE
Volker. |
|
|
| Author |
RE: Network issues with PersonalAlpha container |
gremlin
Member
Posts: 22
Joined: 19.04.10 |
| Posted on May 21 2010 03:38 |
|
|
Hi
Perhaps this could be included in some FAQs/instructions for those converting from PA - there may be many places where networking changes and updates are required?
I think, however, it has made no difference to the underlying network problem. After I rebooted, I still get this showing....
jumping to bootstrap code
OpenVMS (TM) Alpha Operating System, Version V8.3
© Copyright 1976-2007 Hewlett-Packard Development Company, L.P.
%DECnet-I-LOADED, network base image loaded, version = 05.13.00
%SYSINIT-I- waiting to form or join an OpenVMS Cluster
%VMScluster-I-LOADSECDB, loading the cluster security database
%EWA0, FastFD mode set by console
%EWA0, Link state: UP
%MSCPLOAD-I-CONFIGSCAN, enabled automatic disk serving
%CNXMAN, Proposing formation of a VMScluster
%CNXMAN, Now a VMScluster member -- system GAMMA
%CNXMAN, Completing VMScluster state transition
%STDRV-I-STARTUP, OpenVMS startup begun at 21-MAY-2010 23:26:14.06
No communication with other cluster members across the EWA0 link....
and...
$ mc ncp sho exe
Node Volatile Summary as of 21-MAY-2010 23:34:28
Executor node = 1.3 (GAMMA)
State = on
Identification = DECnet-OSI for OpenVMS
So, DECnet is running, EWA0 is selected but.....maybe using a PA image is not a good idea? |
|
|
