x86 PXEBOOT hangs

Rob De Langhe <[email protected]> Sun, 17 Jun 2012 23:06:18 +0200
Newsgroups gmane.os.solaris.managers
Message-ID <[email protected]>
hi managers,

I am suffering with the PXEBOOT of a new server. What's happening, is that I see the box getting its IP, hostname "fanny", router, netmask etc via DHCP

Then I get the grub-menu, with the Solaris-10u9 installation option from the Jumpstart installation, which I obviously choose.

Meanwhile I see the TFTP being done, and then a lot of NFS packets. On the screen there are dots being printed, then suddenly nothing anymore, and it stays like that forever. The last packets are about NFS mounts.
The server is named "x2100" :

       fanny -> x2100        length:   60  TCP D=2049 S=1018 Ack=1376342626 Seq=58208830 Len=0 Win=49640
       fanny -> x2100        length:  238  NFS C 4 (secinfo     ) PUTROOTFH LOOKUP spare LOOKUP store LOOKUP jumpstart LOOKUP config SECINFO sysidcfg
       x2100 -> fanny        length:  162  NFS R 4 (secinfo     ) NFS4_OK PUTROOTFH NFS4_OK LOOKUP NFS4_OK LOOKUP NFS4_OK LOOKUP NFS4_OK LOOKUP NFS4_OK SECINFO NFS4_OK AUT...
       fanny -> x2100        length:  342  NFS C 4 (mount       ) PUTROOTFH GETFH LOOKUP spare GETFH GETATTR c8000167 0 LOOKUP store GETFH GETATTR c8000167 0 LOOKU...
       x2100 -> fanny        length:  902  NFS R 4 (mount       ) NFS4_OK PUTROOTFH NFS4_OK GETFH NFS4_OK FH=5F97 LOOKUP NFS4_OK GETFH NFS4_OK FH=62C8 GETATTR NFS4_OK LOOK...
       fanny -> x2100        length:  210  NFS C 4 (fsinfo      ) PUTFH FH=9175 GETATTR 20e00000 1c00
       x2100 -> fanny        length:  190  NFS R 4 (fsinfo      ) NFS4_OK PUTFH NFS4_OK GETATTR NFS4_OK
       fanny -> x2100        length:   60  TCP D=2049 S=1022 Ack=1391073705 Seq=55262629 Len=0 Win=49640
       fanny -> 192.168.1.255 length:  146  BPARAM C GETFILE ns
       fanny -> 192.168.1.255 length:  146  BPARAM C GETFILE ns (retransmit)
       fanny -> x2100        length:   60  TCP D=2049 S=1022 Fin Ack=1391073705 Seq=55262629 Len=0 Win=49640
       x2100 -> fanny        length:   54  TCP D=1022 S=2049 Ack=55262630 Seq=1391073705 Len=0 Win=49640
       x2100 -> fanny        length:   54  TCP D=1022 S=2049 Fin Ack=55262630 Seq=1391073705 Len=0 Win=49640
       fanny -> x2100        length:   60  TCP D=2049 S=1022 Ack=1391073706 Seq=55262630 Len=0 Win=49640
       fanny -> x2100        length:  342  DHCP/BOOTP DHCPREQUEST
       x2100 -> fanny        length:  543  DHCP/BOOTP DHCPACK
       fanny -> x2100        length:  342  DHCP/BOOTP DHCPREQUEST
       x2100 -> fanny        length:  543  DHCP/BOOTP DHCPACK

The contents of my /tftpboot/boot/grub/menu.lst :

 menu.lst
default=0
timeout=30
min_mem64 1024
title Solaris_10 install10u9_x86
        kernel$ /I86PC.Solaris_10-1/multiboot kernel/amd64/unix -v -B install_debug=enable,input-device=ttya,output-device=ttya,ttya-mode="115200,8,n,1,-",install_config=192.168.1.19:/spare/store/jumpstart/config,sysid_config=192.168.1.19:/spare/store/jumpstart/config/sysidcfg,install_media=192.168.1.19:/spare/store/jumpstart/install10u9_x86
        module$ /I86PC.Solaris_10-1/amd64/x86.miniroot

I added the "install_debug" flag, but I have no idea what that should have generated as debug output, I didn't see anything on the screen.

The "snoop" details tell me that these last NFS packets contain MOUNT requests from the new host "fanny" to Jumpstart server "x2100" for /spare/store/jumpstart/config/sysidcfg.
The reply from "x2100" is some NFS4_OK, but hard to interprete these packets exactly: did the mount happen ok ? On the server "x2100" I have checked with "showmount -a", but that gives me no mounts done by clients:
# showmount -a
#

However the underlying directory is properly shared by the server "x2100" :
# share
-               /spare/store/jumpstart   ro   ""

And the NFS-mount is enabled:
# svcs nfs/server
STATE          STIME    FMRI
online         Aug_04   svc:/network/nfs/server:default

The file "sysidcfg" is correctly there, with the following contents:
# ls -l /spare/store/jumpstart/config/sysidcfg
-rw-rw-r--   1 root   other    175 Jan 27  2005 /spare/store/jumpstart/config/sysidcfg
# cat /spare/store/jumpstart/config/sysidcfg
system_locale=en_US
timezone=MET
timeserver=localhost
terminal=vt100
security_policy=NONE
network_interface=primary
        { dhcp
          protocol_ipv6=no }
root_password=... (hidden)

Does someone have any clue about what might be failing ? If not, can someone give me suggestions about how I could get more verbosity on the screen to help me identifying the steps that is failing ?

many thanks in advance !
Rob



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.