Hi team,
I have just seen the new doc wrt otherinterfaces to be deprecated, and a lot of our customers we prefix with the net name typically similar to below
ibnode001
bmcnode001
etc...
So would like the ability to prefix rather than suffix the normal hostname.
But I like the idea though overall
Norm, let's talk about this
Ran out of time for 2.8.2
Also - this would be new support and we should schedule FVT.
old request - I won't have time to get to it at this point - not sure it is needed?
Yeah, an old request, that I am still keen on.
If you guys don't have the resources, I will try to spend time to see if I can implement it myself, and forward on the patches
As for a lot of my customers we have a prefix rather than suffix
Hi,
Below is my first attempt
http://sourceforge.net/u/arifali/xcat/ci/e01827f9d3f78dedf4ef39d4c72a8fdd1bb5c3b5/
if this looks good, I can submit a merge request
Any comments/sauggestions are welcome
regards,
Arif
Thanks we will review and checkin
Looking at your change. I think you also need to add the new nichostnameprefixes attribute to the defspec section of the Schema.pm file for the *def comands to work.
nics table
Arif with this update, go ahead and check the code into the 2.9 ( master) git stream. I will add the testing to out 2.9 schedules.
As requested, I have also added nodeprefix to @nodeattrs, but there is no mention of any of the nics for suffix in defspec, so I presume, both the suffix and prefix need to be added to defspec?
checked-in into master, commit dd3f08
yes they both should be in there.
right, I will get those added and tested
defspec has it, but lsdef doesn't show them,
should they?
yes it should but I see my entry looks like below.
We must be dealing with some special code in the lsdef to output more than just what is in the attribute . Will search for it.
nichostnamesuffixes.eth1=-eth1|-eth1:1
nicips.eth0=172.20.101.2
nicips.ib0=168.1.0.2
nicips.ib1=168.2.0.2
nicnetworks.eth1=60_0_0_0-255_255_255_0|60_0_0_0-255_255_255_0
nictypes.eth1=ethernet|ethernet
I found one minor problem in DBobjectdefs.pm and one big one
But the big problem is the use of DBobjUtils->expandnicsattr to create the output for lsdef. It expects the format of the nichostnamesuffix in DBobjectdefs.pm. In each of the places it is called we are going to have to write a new routine that will format the nichostnameprefix output and check which attribute we are processing.
Would you do that. I checkin my small fix
Thanks Lissa for looking,
I have looked through, and can't see anything wrong. What is the wrong output you would get, and what is the expected output. (I don't know the code as well as you guys). Below are some output from some of the commands, and all seems to be working, after installing the latest snapshot
On another note, I was having issues with
makehosts -a
, where it was adding another entry in\etc\hosts
due to the fact the domain variable had an extra space; revert a problem I created in a previous commit. The following fixes both problems.https://gitlab.arif-ali.co.uk/arif/xcat-core/commit/4f2bce8c28ef76d26346f7f56415661fab987ff3
Last edit: Arif Ali 2014-01-13
nichostnameprefixes should show up when you run lsdef <nodename>. It does not; although nichostnamesuffixes does. That is because of that formatting routine. If you don't care we can fix it later and I will leave the defect open.
tabdump nics
node,nicips,nichostnamesuffixes,nichostnameprefixes,nictypes,niccustomscripts,nicnetworks,nicaliases,comments,disable
"cn14",,"eth1!-eth1|-eth1:1","ib",,,,,,
Object name: cn14
arch=ppc64
cons=hmc
currstate=netboot rhels6.4-ppc64-compute
groups=lpar,all
hcp=c76v2hmc01
hwtype=lpar
id=34
initrd=xcat/osimage/rhels6.4-ppc64-netboot-compute/initrd-stateless.gz
installnic=mac
kcmdline=imgurl=http://!myipfn!:80//install/netboot/rhels6.4/ppc64/compute/rootimg.gz XCAT=!myipfn!:3001 NODE=cn14 FC=0 BOOTIF=e6:ef:63:d5:f0:02
kernel=xcat/osimage/rhels6.4-ppc64-netboot-compute/kernel
mac=e6:ef:63:d5:f0:02
mgt=hmc
netboot=yaboot
nichostnamesuffixes.eth1=-eth1|-eth1:1
nodetype=ppc,osi
os=rhels6.4
parent=Server-9125-F2A-SN027ACB4
postbootscripts=otherpkgs
postscripts=syslog,remoteshell,syncfiles
pprofile=compute
primarynic=mac
profile=compute
provmethod=rhels6.4-ppc64-netboot-compute
status=booted
statustime=07-31-2013 08:21:27
updatestatus=failed
updatestatustime=08-13-2013 12:29:16
Could you please take the makehosts fix and open another defect and attach it. I need the defect to track what we do. I will review it and then if ok you can checkin. We appreciate the help.
Last edit: Lissa Valletta 2014-01-16
OK, I see the problem now, I will have a look tomorrow, and will get the other defect open as well