Conference aosg::lsm

798.0. "LSM root recovery" by VAXRIO::LEO () Wed May 28 1997 12:17

    Hi,
    
    	I am using Digital Unix 4.0b.
    
    	My customer wants to implement an contingency plan in case he looses 
    	some files from the / (root) file system. Not all the / file system.
    
    	It has already happened some times.
    
    	My customer don't want to go through all the steps below, it means :
    
    	1) boot from CD-ROM
    	2) create root_domain and filesets
    	3) mount root file system
    	4) vrestore
    	5) reboot
    	6) volencap of the root file system
    	7) 2 reboots generated by volencap
    
    	It can take more than 15 minutes. It is too much for him.
    	 
    	I am trying to implement an alternative way to do that.
    
    	The question is still the same :
    
    	How can I recover any file that was placed on the root file system 
    considering that it is  an LSM encapsulated file system ?
    
    	New alternative :
    +++++++++++++++++++++++
    
    	I have rebooted the Operating system using the Digital
    Unix CD-ROM.
    
    	After that what I am trying to do is just to follow the instructions 
    shown on pages 7-14 and 7-15 of the Digital Unix 4.0B "Installation
    Guide" in order to rebuild the LSM configuration : 
    
    	But I am not able to execute the step number 4 :
    
    	4 ) # vold -m disable
    
    	The error message that I get is :
    
       "LSM: vold: error: enable failed: cannot create portal 
    	Read only file system; aborting"
    
    	I don't remember this error message on Digital Unix v3.2x.
    
    	After that I have tried an workaround to bypass this message :
    
       #  mkdir /var/vol (/var is already an mfs file system)
       #  vold -R /var/vol -D /var/vol -m disable     
        
    	   This command worked fine.
    
    	So I have tried to do the step number 5:
    
       # voldctl init <name-of-the-host>
    	
    	I got the following error message :
    
    	voldctl: system error: no such file or directory
    
    	I am wondering that I should inform the new directory /var/vol to
    the voldctl init command. How can I do that ?
    
    	I think if I can solve this problem I will be able to continue
    the recovery steps.
    
    	If I can issue all the 11 steps (using already defined scripts) I
    think I can reduce the downtime to less than 3 minutes. I will be able
    just to restore the specifics file that were lost.
    
    
    	Any help would be very appreciated.
    
    	Thanks in advance.
    
    	Best regards,
    
    	Leo
    	Digital Technical Support
    	Brazil
    	
    
    
    
    
    
    
    		
    	
    
    
        I have already done all these steps without any error using Digital
    Unix version 3.2x.
    	
    	
    
     

The rendezvous file pathname option (-R) is neither supported nor documented
for the other commands (note, leaving a reference to it within the vold(8) 
man page was an oversight), but the code's there and you can give it try.  
For example:
    # mkdir /var/vol
    # vold -r reset -m disable -R/var/vol/vold_request -D/var/vol/vold_diag
    # voldctl -R/var/vol/vold_diag init [hostid]
    # voldctl -R/var/vol/vold_diag mode
    # voldctl -R/var/vol/vold_diag enable

    Once the volume daemon is enabled, you should be able to run the 
    commands with -R/var/vol/vold_request added as an option.

Once you worked out the rest, post the entire list of steps if you want
some more feedback.  

BTW, if the steps documented in the Install Guide no longer work, please
file a QAR.

thanks,
--steve

[Posted by WWW Notes gateway]

    Leo,
    
 >>    "LSM: vold: error: enable failed: cannot create portal 
 >>          Read only file system; aborting"
    
    
    Did you make the root filesystem writeable ? (mount -u /)
    
    Regards, Manu.