ELC: Can't retain IP, Can't Ping itself or others, RARP & rpc.lockd errors
The cause of my problem is that the existing OS configuration is
confused. Parts of the OS things its trying to use either NIS or DNS or
both. I do not want the OS to use either.
The best fix came from:
Bismarks recommended was forcing the ifconfig configuration at bootup
via changing entries in /etc/rc.local file.
The existing entry:
ifconfig -a netmask + broadcast + > /dev/null
was changed to:
ifconfig le0 %.%.%.% netmask 255.255.255.0 broadcast %.%.%.255 >
/dev/null
### %.%.%.% = IP address - not given for security reseaons.
### %.%.%.255 = first three fields of IP address + .255
And this works. I still get error messages from /etc/rc.boot but the
machine connects to the network, is ping-able, is rlogin-able, and there
are no errors in the system console concerning rpc.lockd. In short it
works! Thank you Bismark.
Thanks also to:
Ayoub, Nicky
Again, thanks from Chris O+Neal
The original email problem message for help was:
Dear Fellow Sun Managers,
Hi?
I have a problem and am asking for help.
I have a SPARCstation ELC not retaining its IP address for ifconfig,
le0: RARP timed out, rpc.lockd: cannot contact status monitor. And unit
can not ping itself nor be ping+ing by others. Unit otherwise boot and
runs OK.
The ELC is "new to me" and is replacing a "dead" SLC which booted it
unix (Sun OS 4.1.3_U1) from an existing external hard disk connected
directly to the unit. I have just recently inherited the SYS ADMIN
responsibilities for this department and am unfamiliar as to how the SLC
was original setup. I do not want NIS or DNS running.
The ELC booted immediately with user crons and software once existing
external hard drive was attached but failed to make network connection.
The displayed errors are:
#A.) Displayed errors from /etc/rc.boot at bootup are:
ifconfig: ioctl (SIOCSIFDSTADDR): Invalid argument
le0: RARP timed out.
#B.) Displayed error from /etc/rc.local at bootup are:
le0: flags=43<UP, BROADCAST, RUNNING>
inet 0.0.0.0 netmask ff000000
ether 8:0:20:3:b0:37
lo0: flags=49 <UP,LOOPBACK,RUNNING>
inet 127.0.0.1 netmask ff000000
#C.) Displayed error from console window are:
rpc.lockd: Cannot contact status monitor!
I am really stumped. Previous to emailing you I have done my own
independent research which resulted in the 14 steps I have already taken
below. The problem remains +un-solved+ and I enclose +what I have done+
in the hopes that it will lite the lightbulb in another mind as to what
I have to do to get this ELC to network. Things I have done since the
ELC boots but does not attached to network are:
#1.) NVRAM setting were verified at ok prompt:
+printenv diag-switch?+
Results: Returns +False+
+printenv boot-device+
Results: Returns +disk+
#2.) Boot prompt network hardware testing was done at ok prompt:
+test net+
Results: Both internal and external loopback tests successfully passed.
+watch-net-all+
Results: +?+ I guess "watch-net-all" is a more recent command and is not
in my older boot-prom or I don+t know how or were to run it from.
#3.) ELC was connected directly into switched-hub using different patch
cable and tranceiver:
Results: No change.
#4.) Re-set machine for first time install initialization:
+cd /usr/etc/install;sys-unconfig;sync;sync;sync;reboot+
Results: ELC reboots and ask for config info (hostname, Time zone, IP,
NIS). Unit reminders IP and is pingable until first reboot were above
error messages are displayed and network connectivity is lost. No
changes to /etc/hosts or other config files are made in the interim.
#5.) Verified singular existence and integrity of /etc/hostname.le0:
+Ls /etc/hostname.*+
Results: Only one file (hostname.le0) matched above conditions. There
are no others like; /etc/hostname.le0~ or like; /etc/hostname.old etc?
+vi /etc/hostname.le0+
+:set list+
Results: Single word (hostname) on first line of file with no control
characters other than +$+ for end-of-line.
#6.) Reviewed the integrity of /etc/hosts:
+more /etc/hosts+
Results:
127.0.0.1 localhost
%.%.%.% %%%% loghost ### %.%.%.% = IP address and %%%% = hostname,
both not given for security reasons.
Results: Looks ok to me. Did spot check of original /etc/hosts file. It
has 200 entries but has 127.0.0.1 commented out at top of file. Using
neither the original nor setups new host file enabled ELC to connect to
network.
+vi /etc/hosts+
+:set list+
Results: No control character other than +$+ and +^I+ were displayed.
#7.) Verified existence of /etc/ethers:
+ls /etc/ethers+
Results: No such file exist. My understanding is I don+t really need
this file, and its better not to have one than to have one that is
mis-configured. Is this true?
#8.) Verified ifconfig configuration call in /etc/rc.local:
+vi /etc/rc.local+
+/ifconfig+
Results: Found two lines; ifconfig -a netmask + broadcast + > /dev/null,
and ifconfig -a. Everything looked ok to me, anything else I should be
looking for?
#9.) Verified content of /etc/netmasks:
+more /etc/netmasks+
%.% 255.255.255.0 ### %.% = first two fields of ELC IP
address.
Results: There were 18 other line entries, the above is the first line
in the file. I do not recognizes the 17 netmasks but that does not mean
that they are not here on campus.
#10.) Verified that /etc/sm and /etc/sm.bak were clean:
+\rm -R /etc/sm /etc/sm.bak+
Results: Problems continued but after each reboot these directories
re-made.
#11.) Manually reconfigured ifconfig:
+ifconfig le0 down; ifconfig le0 %.%.%.%+
Results: ELC could ping itself but not other network machines. ELC lost
manually setting after reboot.
#12.) Verified existence / non-existence of DNS:
+ls /etc/resolv*+
Results: /etc/resolve.conf.save was returned. I believe that this file
was made by sys-unconfig done in step #4 above, but I am not sure. So
does this mean that DNS was running on this machine? I don+t know. I
could not find /usr/lib/libc.so.1.8.1, there was a /usr/lib/libc.so.1.9
and I did not deleted it. How would I uninstall DNS if it were loaded?
#13.) Verified existence / non-existence of NIS:
`more /etc/defaultdomain+
Results: Returned single line with single word +noname+. I removed this
file.
+cd /var/yp/binding+
Results: Directories; /var/yp/* do not exist.
#14.) Verified existence / non-existence of gateway config file:
`more /etc/gateway+
Results: No such file exist.
