You need to include the filesets reported as "missing". A SPOT (shared product object tree) is like a shared, NFS-mounted "/usr"-tree and therefore needs several core packages included otherwise it couldn't act in this function. If you do not have them in your LPP-source you should include them there either.
Further, this is only a suspicion but it seems you manipulated the exports file yourself. As a rule of thumb: don't do that! At least don't do that with anything in or below the directory with the NIM resources. It might solve one specific problems but usually introduces several dozens of new ones in the process.
Another frequent problem is the definition of the NIM networks. NIM works (in parts) on top of the bootp protocol and therefore has to specifically check clients on other networks if they can be booted with it (this means testing the switch/router for being able to act as a "bootp relay agent", originally called "bootp forwarding agents", see RFC951, RFC1532, RFC1542, etc.) which is why for each network a client could be on a "NIM network" has to be defined.
Finally, when you allocate a resource to a client and it doesn't work somehow in most cases - but not all! - the resources are deallocated. Alas, rather frequently they remain allocated and not even a deallocation with the "force" option will straighten them out. NIM is quite untrustworthy in this regard and i hate to say it but it didn't improve much over the ~20 years i work with it. It is sometimes a good idea to simply delete the respective machines object and recreate it from scratch.
Right now i am not at the office so i cannot look it up but tomorrow i can post you a simpler way to create SPOTs, so stay tuned.
This error happened on my first try. So as part of common sense troubleshooting:
-I unexported everything with exportfs -ua
-Backuped the /etc/exports
-modified the exportfs file to add directly the folder where the SPOT is.
-Exported everything with exportfs -a
Once I noticed the error was still occuring I figured the Folders added in /etc/exportfs are reccursive, therefore I rollbacked.
I'm looking foward to see your procedure
Trivium012
---------- Post updated at 10:59 AM ---------- Previous update was at 10:22 AM ----------
So I decided to try with a mksysb instead and I found a mksysb ressource that was made prior a prior sysadmin. So I made the SPOT with it, but the error messahe occured again... but this time, when I list the new SPOT, it says Ready for use
I will try with this and keep you posted!
Also, if it works, I also want to know how to fix the problem above! For my own understanding (I wish I could simulate a PowerPC at home for my personal tests )
---------- Post updated at 11:33 AM ---------- Previous update was at 10:59 AM ----------
Ok so I allocated my SPOT and my MKSYSB to my new LPAR and started the software installation. But when the LPAR boots, I get stuck on error 0611 :Remote mount of the NFS file system failed.
But all my nfs daemons seems to be running fine (I think?)
First off, i want to apologize. Having a very busy day at the office made me forget to update this thread, i promise to make good for that first thing tomorrow morning. Sorry.
Quote:
Originally Posted by trivium012
-modified the exportfs file to add directly the folder where the SPOT is.
-Exported everything with exportfs -a
This is probably where your error is. You DO NOT EXPORT ANYTHING! Instead "allocate" a resource (SPOT, mksysb, ...) to a client. NIM will do the exporting, mounting, unmounting and unexporting all on its own. You "allocate" (NIMs wording) the resource using the "nim" command:
or you can do it using a SMIT panel (search in "NIM->NIM Administration Tasks->somewhere", i don't know it by heart).
You might want to read this little introduction i once wrote in the meanwhile to get acquainted to NIMs and AIXs terminology.
There are, of course, more in-depth sources too: i can wholeheartedly recommend the IBM Redbook on NIM. Despite the seemingly outdated version number it is still quite up-to-date. NIM is a very stable software and major changes haven't taken place since this book.
Quote:
Originally Posted by trivium012
But all my nfs daemons seems to be running fine (I think?)
Yes, if you use NFS3 or lower.
If you use NFS4 you need nfsrgyd to be running. Further you need to have both machines (server and client) in the same NFS domain. Use the chnfsdom command to display and/or set the NFS domain. I use to set it to a value equal to the respective DNS domain, but this isn't required. Notice that for NFS4 you also need some FS as a "stable storage", preferably one without an external JFS log but using an "inline log" instead.
OK, sorry for the delay, here is my approach to managing NIM resources:
Start with an LPP_SOURCE. To avoid having update packages without the respective base-level filesets in my LPP_SOURCES i always copy the base-level LPP_SOURCE (which contains only install-packages and no update-packages) first. Then i create version X+1 using the resource of version X as a basis. Here is how i created the AIX 7.1 TL3 SP3 LPP_source from the 7.1 TL2.
First, copy the LPP_SOURCE of your base level fileset:
after this you update the newly created image(-copy) with your updates to the level you want. /nimres/aix_71/lpp_71_03_source holds the packages i downloaded from fix-central:
Once this is through i create the NIM-resource to use that image:
To remove all the unnecessary languages you don't need (perhaps you need only "C", "en_US" and maybe your local locale):
Now your LPP_source is ready to use. If you need a SPOT you create it from the LPP-SOURCE:
Here is how you install packages through NIM:
a) from the NIM-Master:
b) from the NIM-client (the /etc/niminfo file has to be configured already and the client needs to be configured on the NIM-master) :
I hope this helps.
So while you answered me, I finally managed to install the OS by making a spot from a mksysb that was already there. So main problem fixed. But I still want to be able to create my own lpp_source and SPOT (for my own knowledge and to document properly)
So I downloaded the base image of AIX from the IBM website and created an lpp_source from it.
A crap load like the last lines...
So I'm guessing I really do have an NFS problem, but I have no idea how to troubleshoot this problem...
Here's my exports file
Thanks for the help!!
Moderator's Comments:
edit by bakunin: i took 20 min out of my life to remove (partially) all the superfluous markup. Sorry, i give up.
Hi all
I am facing difficulties when using the useraddcommand.
when running useradd -u 60006 -g 100 -d /opt/xxx/yyy/z -m -s /bin/false ftpuser
It does create the user, but with a wrong home directory. It creates its home directory at /home/ftpuser
Please can you help? (2 Replies)
Discussion started by: fretagi
2 Replies
2. Forum Support Area for Unregistered Users & Account Problems
Hi.
I registered, but when trying to post, I get ...
"Sorry, you do not have enough Bits to post a new thread"
Google & Forum search showed some hits, but no solution.
Please advide.
Thanks! (1 Reply)
Hello All,
Kindly need help in my problem i.e. I get this error when i create spot from mksysb.
#0042-001 nim: processing error encountered on "master":
0042-001 m_mkspot: processing error encountered on "master":
0042-175 c_mkspot: An unexpected result was returned by the... (2 Replies)
Hi,
I am unable to create/open any file in vi editor for normal user, though using root I am able to create/open any file, using redhat 5.5,
example
vi test
it showing nothing after entering command (9 Replies)
Hi All,
I am trying to create a folder in /export/home/user1 directory as a user1 but I am getting the error message as
-bash-3.00$ mkdir abc
mkdir: Failed to make directory "abc"; Permission denied
Also I tried creating the folder but it fails too
-bash-3.00$ touch abc
touch: cannot... (2 Replies)
Hi,
I have a process which creates pthreads to generate some reports. After creating the reports these threads return null. But after 1024 threads, the process is not able to create any threads further.,and at max 5 threads are existing simultaneously and are returning the control back after... (2 Replies)
Good morning, I have no unix training. My daily backup failed with the following error. ' Cannot create /tmp/tape.bootimage because it already exists. I am using smit to run a rootvg and SAPvg backup. What should I do? (1 Reply)
I am using Solaris 9 machine. I am unable to create any aliases.
I gave the following command:
alias ls='ls -ltr'
It did nt give any errors. But next time, I use ls, it is working simply as ls but not like ls -ltr.
I am not having admin permissions. Is it the reason. (3 Replies)
Hello friend,
Now i'm doing JDS3.1 in v440 servers.while checking the lustatus command
getiing out like this
lustatus
Boot Environment Is Active Active Can Copy
Name Complete Now On Reboot Delete Status
-------------------------- --------... (0 Replies)
Hi,
I'm trying to create a new group on an AIX 5.2.0.0. server but am getting the following error:
root:/root $ mkgroup debt
3004-698 Error committing changes to "debt" : Value is invalid.
I've change the group name to no affect.
Does anyone have any ideas what could be causing this?
... (0 Replies)