This is an odd situation, perhaps someone else has run into it..
I've been running jumpstart off of the netapp, Solaris-2.6 and -8.
For both versions, and for any network configuration, any time I netboot a system off of the netapp, I must send one ping packet to the system being jumpstarted to get it beyond "configuring devices..." (Sol-2.6) or "Configured interface hme0" (Sol-8, ultrasparc, same effect for sun4m, just s/hme/le/). This ping packet must be directed specifically at the host, not a broadcast ping ("ping a.b.c.255").
Anyone else seen this? Is there a workaround or hack that makes this go, other than what I'm doing already?
Rob++ ---------------------------------------- Rob Windsor E-Mail - mailto:rob_windsor@adc.com Senior Unix Systems Administrator Voice - phone:972-680-6919 Computer Services Fax - phone:972-680-0370 Broadband Access and Transport Group __o ADC Telecommunications _`<,_ Richardson, TX 75082 (_)/ (_)
Don't know if this is your answer, but I have this note from our jumpstart server:
- 5.[678]*/Solaris_*/Tools/Boot/sbin/rcS modified to not run get_netmask
Here's the diff ... < if [ -x /no/sbin/get_netmask ]; then ---
if [ -x /sbin/get_netmask ]; then
I never tried sending a ping, so I'm not 100% sure if this is the same issue you're having, but without this it would hang "indefinately" when a netapp is used as the fileserver.
Hope that helps.
Bryan
On Thu, Dec 07, 2000 at 02:15:09PM -0600, Rob Windsor wrote:
This is an odd situation, perhaps someone else has run into it..
I've been running jumpstart off of the netapp, Solaris-2.6 and -8.
For both versions, and for any network configuration, any time I netboot a system off of the netapp, I must send one ping packet to the system being jumpstarted to get it beyond "configuring devices..." (Sol-2.6) or "Configured interface hme0" (Sol-8, ultrasparc, same effect for sun4m, just s/hme/le/). This ping packet must be directed specifically at the host, not a broadcast ping ("ping a.b.c.255").
Anyone else seen this? Is there a workaround or hack that makes this go, other than what I'm doing already?
Rob++
Rob Windsor E-Mail - mailto:rob_windsor@adc.com Senior Unix Systems Administrator Voice - phone:972-680-6919 Computer Services Fax - phone:972-680-0370 Broadband Access and Transport Group __o ADC Telecommunications _`<,_ Richardson, TX 75082 (_)/ (_)
On Thu, 07 Dec 2000 14:55:58 PST, Bryan Stansell wrote:
Don't know if this is your answer, but I have this note from our jumpstart server:
- 5.[678]*/Solaris_*/Tools/Boot/sbin/rcS modified to not run get_netmask
Here's the diff ...
< if [ -x /no/sbin/get_netmask ]; then
if [ -x /sbin/get_netmask ]; then
I never tried sending a ping, so I'm not 100% sure if this is the same issue you're having, but without this it would hang "indefinately" when a netapp is used as the fileserver.
Hope that helps.
This did the trick!
Thanks.
(Leave it to a fellow former-Synopsys-er to have the answer! :)
Rob++ ---------------------------------------- Rob Windsor E-Mail - mailto:rob_windsor@adc.com Senior Unix Systems Administrator Voice - phone:972-680-6919 Computer Services Fax - phone:972-680-0370 Broadband Access and Transport Group __o ADC Telecommunications _`<,_ Richardson, TX 75082 (_)/ (_)
Looking up 'get_netmask' in Sun's SunSolve database, the problem is lack of response to an ICMP Address Mask Request. (Which NetApp probably does not support in ONTAP.)
Bryan Stansell wrote:
Don't know if this is your answer, but I have this note from our jumpstart server:
5.[678]*/Solaris_*/Tools/Boot/sbin/rcS modified to not run get_netmask
Here's the diff ...
< if [ -x /no/sbin/get_netmask ]; then
if [ -x /sbin/get_netmask ]; thenI never tried sending a ping, so I'm not 100% sure if this is the same issue you're having, but without this it would hang "indefinately" when a netapp is used as the fileserver.
Hope that helps.
Bryan
On Thu, Dec 07, 2000 at 02:15:09PM -0600, Rob Windsor wrote:
This is an odd situation, perhaps someone else has run into it..
I've been running jumpstart off of the netapp, Solaris-2.6 and -8.
For both versions, and for any network configuration, any time I netboot a system off of the netapp, I must send one ping packet to the system being jumpstarted to get it beyond "configuring devices..." (Sol-2.6) or "Configured interface hme0" (Sol-8, ultrasparc, same effect for sun4m, just s/hme/le/). This ping packet must be directed specifically at the host, not a broadcast ping ("ping a.b.c.255").
Anyone else seen this? Is there a workaround or hack that makes this go, other than what I'm doing already?
Rob++
Rob Windsor E-Mail - mailto:rob_windsor@adc.com Senior Unix Systems Administrator Voice - phone:972-680-6919 Computer Services Fax - phone:972-680-0370 Broadband Access and Transport Group __o ADC Telecommunications _`<,_ Richardson, TX 75082 (_)/ (_)
Anyone else seen this? Is there a workaround or hack that makes this go, other than what I'm doing already?
Sun has a bug ID on this issue: 4218092, dated 14 Mar 1999. What's going on here is that that the Sun client runs /sbin/get_netmask from /etc/rcS (in the boot image). get_netmask sends an ICMP Address Mask Request to the boot server. The filer doesn't respond and get_netmask never times out, nor does it issue an explanatory message. I'm not clear on why the ping unfreezes things.
When I ran into this I came up with several possible workarounds, including
* Use a Sun box as the boot server. The install server can still be a filer.
* Hardcode the netmask into /etc/rcS and dispense with get_netmask.
* Hardcode the address of a Sun box into the get_netmask call, rather than using the boot server.
The first approach has the advantage of not having to modify the boot image. Of course, if Data ONTAP responded to address mask requests this would all be moot.
Jeff
-
Bryan Stansell -
Jeff Gilliam -
Matthew Stier -
Rob Windsor -
Rob Windsor