On 9/27/10 8:46 AM, Dennis Stam wrote:
> Greg,
>
> Thanks for pointing that out. That could be the problem.
>
> By default SALI does not wait for fetching a DHCP lease. But you can add
> the variable ETHER_SLEEP.
>
> https://subtrac.sara.nl/oss/sali/wiki/SaliUsage/Installationparameters#ETHER_SLEEPsec
>
>
> Regards,
>
> Dennis
>
> On 2010-09-27 14:20, Fiumara, Gregory wrote:
>>
>> On Sep 25, 2010, at 10:14 AM, Chris Pepper wrote:
>>> No, it's getting a lease. Otherwise it wouldn't be able to download the
>>> code and wouldn't be trying to ping IMAGESERVER, but I do see it in
>>> the log.
>>
>> I have seen cases with clients getting a DHCP lease at the start of
>> the PXE boot, but when the SystemImager scripts start running and
>> request another DHCP lease, the request fails. In that case, the nodes
>> were connected to a newly configured Cisco switch that had spanning
>> tree enabled, which imposed a delay, preventing a lease, and
>> ultimately making the ping test fail.
>>
>> Not saying that's your situation, but I have seen it.
Thanks. I added "ETHER_SLEEP=30" to
/var/lib/tftpboot/pxelinux.cfg/default in case that helps, but I don't
think that's my problem -- I'd expect a failure to connect rather than a
syntax error if it was giving the correct argument, and I was already
seeing the DHCP server hand out a lease again after SALI's components
are fetched.
Chris
Without ETHER_SLEEP:
> Sep 27 11:55:39 lux dhcpd: DHCPDISCOVER from 1c:c1:de:f0:8d:88 via eth1
> Sep 27 11:55:39 lux dhcpd: DHCPOFFER on 10.0.2.114 to
1c:c1:de:f0:8d:88 via eth1
> Sep 27 11:55:41 lux dhcpd: DHCPREQUEST for 10.0.2.114 (10.0.2.10)
from 1c:c1:de:f0:8d:88 via eth1
> Sep 27 11:55:41 lux dhcpd: DHCPACK on 10.0.2.114 to 1c:c1:de:f0:8d:88
via eth1
> Sep 27 11:55:41 lux xinetd[6320]: START: tftp pid=13135 from=10.0.2.114
> Sep 27 11:55:41 lux in.tftpd[13136]: RRQ from 10.0.2.114 filename
pxelinux.bin
> Sep 27 11:55:41 lux in.tftpd[13136]: tftp: client does not accept options
> Sep 27 12:05:17 lux in.tftpd[13286]: RRQ from 10.0.2.114 filename
pxelinux.bin
> Sep 27 12:05:17 lux in.tftpd[13287]: RRQ from 10.0.2.114 filename
pxelinux.cfg/01-1c-c1-de-f0-8d-88
> Sep 27 12:05:17 lux in.tftpd[13288]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A000272
> Sep 27 12:05:17 lux in.tftpd[13289]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A00027
> Sep 27 12:05:17 lux in.tftpd[13290]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A0002
> Sep 27 12:05:17 lux in.tftpd[13291]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A000
> Sep 27 12:05:17 lux in.tftpd[13292]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A00
> Sep 27 12:05:17 lux in.tftpd[13293]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A0
> Sep 27 12:05:17 lux in.tftpd[13294]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0A
> Sep 27 12:05:17 lux in.tftpd[13295]: RRQ from 10.0.2.114 filename
pxelinux.cfg/0
> Sep 27 12:05:17 lux in.tftpd[13296]: RRQ from 10.0.2.114 filename
pxelinux.cfg/default
> Sep 27 12:05:17 lux in.tftpd[13297]: RRQ from 10.0.2.114 filename
sali-kernel
> Sep 27 12:05:17 lux in.tftpd[13298]: RRQ from 10.0.2.114 filename
/sali-initrd
> Sep 27 12:05:45 lux dhcpd: uid lease 10.0.2.199 for client
1c:c1:de:f0:8d:88 is duplicate on 10.0.2/24
> Sep 27 12:05:45 lux dhcpd: DHCPDISCOVER from 1c:c1:de:f0:8d:88 via eth1
> Sep 27 12:05:45 lux dhcpd: DHCPOFFER on 10.0.2.114 to
1c:c1:de:f0:8d:88 via eth1
> Sep 27 12:05:45 lux dhcpd: uid lease 10.0.2.199 for client
1c:c1:de:f0:8d:88 is duplicate on 10.0.2/24
> Sep 27 12:05:45 lux dhcpd: DHCPREQUEST for 10.0.2.114 (10.0.2.10)
from 1c:c1:de:f0:8d:88 via eth1
> Sep 27 12:05:45 lux dhcpd: DHCPACK on 10.0.2.114 to 1c:c1:de:f0:8d:88
via eth1
ETHER_SLEEP added & SALI updated:
> [r...@lux tftpboot]# tail -4 pxelinux.cfg/default
> label sali
> kernel sali-kernel
> append load_ramdisk=1 prompt_ramdisk=0 initrd=/sali-initrd
root=/dev/ram rw ramdisk_size=103819 IMAGESERVER=10.0.2.10
ETHER_SLEEP=30 SCRIPTNAME=l01.201009.master
> # console=ttyS0,9600n8
> [r...@lux tftpboot]# md5sum sali-*
> e9267c3b22ce40008ea3f3a644cfdfc9 sali-initrd
> 82909d598ec8de55f6feb2e679ab858d sali-kernel
But apparently IMAGESERVER still isn't being passed to the ping
command.
Chris
------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
sisuite-users mailing list
sisuite-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sisuite-users
