Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Stopped at final stages of netbooting MicroVax II with NetBSD

Well, out of respect for Johnny's wisdom on linux., I gave up on that attempt.  Then I figured - well, mop is supposed to work on VMS, but could not get anything working using the DEC Alpha server - a member of the comp.os.vms group tried it and found he had to make so many patches it was not worth it.

So I thought I would build a MicroVax II under SimH and load NetBSD onto it. (Latest from their web site.) I followed the NetBSD how-to.

Guess what!  I get exactly the same panic dump after it loads using NFS:

..........   1.2100070] rx0 at mscpbus2 drive 1: RX33                                                         
[   2.5700070] boot device: qe0                                                                        
[   2.6100070] root on qe0                                                                             
[   2.6400070] nfs_boot: trying DHCP/BOOTP                                                             
[   5.8700070] nfs_boot: DHCP next-server:                                                
[   5.8800070] nfs_boot: my_domain=ve3id.ddns.net                                                      
[   5.8900070] nfs_boot: my_addr=                                                          
[   5.9000070] nfs_boot: my_mask=                                                         
[   5.9100070] nfs_boot: gateway=                                                           
[  12.3600070] root on                                                   
[  12.4200070] panic: nfs_mountroot: getattr for root                                                  
[  12.4400070] cpu0: Begin traceback...                                                                
[  12.4500070] panic: nfs_mountroot: getattr for root                                                  
[  12.4600070] Stack traceback :                                                                       
[  12.4700070]   Process is executing in user space.                                                   
[  12.4800070] cpu0: End traceback...                                                                  
Stopped in pid 0.1 (system) at  netbsd:vpanic+0x171:    pushl   $0                                     

One member of this group looked at the source code and found that it is trying to getattr on the root directory to get the date and time.

Is it possible that mounting roon on NFS is no longer supported?

If so, does anybody know how far back I would need to go to get a working version?



Amateur Radio, the origin of the open-source concept!
Skype:  TILBURY2591 nw.johnson%ieee.org@localhost

On 2020-12-29 3:03 p.m., Johnny Billquist wrote:
On 2020-12-29 20:09, Nigel Johnson wrote:
I went and set up /etc/exports using yast, thinking that maybe there was
some SuSE-specific magic that was needed, and it now doesn't throw out
my changes and substitute something else!

I think I've already said what I think of Linux... :-)

It is still getting to the point where it tries to mount root after
downloading and panicking with getaddr fail

Should I be worried that there are more than one entry for mountd and nfs?

No. If you pat attention, you'll see that it's for different versions of protocols. So for mountd and nfs, you see both for tcp and udp, and for mountd for version 1, 2 and 3, while for nfs, version 2 and 3.

Which is just fine. I think the network booting and so on, will use version 3 on both, but maybe someone else knows better. But anyway, it's normal to have multiple entries for such things.


T540p:/ # showmount -e
Export list for T540p:
/export/vax/usr  *
/export/vax/root *
T540p:/ # rpcinfo -p
    program vers proto   port  service
     100000    4   tcp    111  portmapper
     100000    3   tcp    111  portmapper
     100000    2   tcp    111  portmapper
     100000    4   udp    111  portmapper
     100000    3   udp    111  portmapper
     100000    2   udp    111  portmapper
     100024    1   udp  59555  status
     100024    1   tcp  38309  status
     100005    1   udp  20048  mountd
     100005    1   tcp  20048  mountd
     100005    2   udp  20048  mountd
     100005    2   tcp  20048  mountd
     100005    3   udp  20048  mountd
     100005    3   tcp  20048  mountd
     100003    2   tcp   2049  nfs
     100003    3   tcp   2049  nfs
     100227    2   tcp   2049  nfs_acl
     100227    3   tcp   2049  nfs_acl
     100003    2   udp   2049  nfs
     100003    3   udp   2049  nfs
     100227    2   udp   2049  nfs_acl
     100227    3   udp   2049  nfs_acl
     100021    1   udp  38005  nlockmgr
     100021    3   udp  38005  nlockmgr
     100021    4   udp  38005  nlockmgr
     100021    1   tcp  43093  nlockmgr
     100021    3   tcp  43093  nlockmgr
     100021    4   tcp  43093  nlockmgr
T540p:/ #


Amateur Radio, the origin of the open-source concept!
Skype:  TILBURY2591 nw.johnson%ieee.org@localhost

On 2020-12-29 10:13 a.m., Johnny Billquist wrote:
By the way, you can verify that things are exported somewhat as
expected by checking with showmount -e
Also, rpcinfo -p is good, to check that both mountd and nfs is
actually registered and running.

If you change /etc/exports, you need to kick mountd (send it a HUP)
for it to actually read in the new exports file.


On 2020-12-29 16:09, Johnny Billquist wrote:
Ok. So the lines before last night essentially gave root no
permissions at all. So that could be a reason for the fail.
Now root and swap are accessible by root as root. usr and home are
not really accessible by root still, but I guess that is fine.

And I assume bonzo is trying to mount /export/client/root as /

So, does it work any better now?


On 2020-12-29 16:04, Nigel Johnson wrote:
The commente-out lines were what wa sin there last night before I added
the four new lines.  bonzo i sthe name of the client.

#/export/vax/root       *(rw)
#/export/vax/usr        *(rw)
#/export/vax/home       *(rw)
/export/client/root -maproot=root:wheel    bonzo
/export/client/swap -maproot=root:wheel    bonzo
/export/client/usr  -maproot=nobody:nobody bonzo
/export/client/home -maproot=nobody:nobody bonzo

Amateur Radio, the origin of the open-source concept!
Skype:  TILBURY2591 nw.johnson%ieee.org@localhost

On 2020-12-28 9:46 p.m., Johnny Billquist wrote:
What does /etc/exports on the exporting host look like?

I hope you are aware that by default, the remote root user gets mapped
to uid -2 on the host...


On 2020-12-28 19:58, Nigel Johnson wrote:
Happy New Year to all,

With lots of help from mind people on this list (remember my
getting mopd to work under linux?), I am now well into netboot.

However, I get to this point and a kernel panic:


[  1.1400070] ra0 at mscpbus0 drive 1:
[   1.2000070] rx0 at mscpbus1 drive 1:
[   2.1700070] ra0: size 650965
[   2.6200070] boot device:
[   2.6600070] root on
[   2.6900070] nfs_boot: trying
[   5.9300070] nfs_boot: DHCP next-server:
[   5.9500070] nfs_boot:
[   5.9800070] nfs_boot:
[   6.0000070] nfs_boot:
[   6.0100070] nfs_boot:
[  12.4200070] root on
[  12.4600070] panic: nfs_mountroot: getattr for
[  12.4900070] cpu0: Begin
[  12.5000070] panic: nfs_mountroot: getattr for
[  12.5100070] Stack traceback
[  12.5300070]   Process is executing in user
[  12.5600070] cpu0: End
Stopped in pid 0.1 (system) at  netbsd:vpanic+0x171:    pushl

The last entry in /var/log/messages on the server is:

2020-12-28T13:51:07.796416-05:00 T540p rpc.mountd[11839]:
mount request from for /export/vax/root

So I am not sure what is going on - can anybody help?

All directories under /export are set RW for root.



Home | Main Index | Thread Index | Old Index