Conference smurf::ase

>    
>    2) Occassionally getting the message "chk_bf_quota: user/group underflow"
>    What does this mean and how do I fix it?
>    

It means that the advfs quota files are not accurate...generally running
vquotacheck (when file systems are quiesent or boot time) will resolve.

sys_check can help give you other clues of things to do with advfs.

I don't know the other answers...

    Hi,
    
    When you create a LSM disk without trailing "c" . I suppose 
    it takes the "g" and "h" partition by default for makeing 
    LSMpub and LSMpriv. 
    			I suppose the disklabel should show you that.
    
    Regards
    Balaji
    

    Well, LSM takes the entire volume. I can check the disklabel, but on a
    4GB disk, I have use of almost the entire disk. My real question is:
    for tie-breakers, can I use rrz136, or must I use an actual partition?
    
    /jim
    

    Well, we have solved questions 2 and 4. Thanks for the help.
    
    We still need help with questions 1 and 3 in the base note. Has anyone
    seen these and have ANY idea what is happening and hopefully how to
    solve them?
    
    Thanks,
    
    /jim
    

Hi

 I suggest your file an IPMT to get the attention you deserve.

    Okay, we seem to have solved question 1. Here is more info on the last
    unanswered question. Before I file an IPMT, maybe somecan tell me what
    is causing this.
    
    2 node Production Server environment: UNIX V4.0B and TCR 1.4. There are
    4 drd services and 2 nfs services. One node boots fine, no problems.
    When the other node boots (this is true if the node is joining the
    cluster, or is the only one coming up), after ASE starts up we get the
    following messages (nodes are mcsteamboat and mctahoe) This ONLY
    happens on mcsteamboat:
    steamboat ASE: mctahoe Agent notice:
    /var/ase/sbin/lsm_dg_action: coldg: Disk group bench01_01_dg: No such
    disk group is imported
    ...voldg deport of disk group bench01_01_dg failed
    ...voldisk: Device rz130: Device is already offline
    
    This repeates for all the disks in the disk group, and for all
    diskgroups and for the nfs services. It appears that mcsteamboat THINKS
    it should own all the services (services are preferred, but they are
    split between the 2 nodes). Once the cluster is up, all works fine. All
    services can fail over, and everything that should work in a cluster
    seems to be working. This appears to be only a startup issue.
    
    Are there any ideas as to why one node would do this and the other node
    is fine? Any suggestions as to how to fix this?
    
    Thanks,
    
    /jim
    

Jim,

What you are seeing is normal. When a DECsafe or TRUcluster (with MC) system
is booting, the stop script of each service is ran (except for the DRD 
services which have no script) and the service is stopped on the booting
system. This operation is done to allow a clean-up of the application(s). 
Indeed, if the system is booting, maybe it is because it has crashed before,
and thus a clean-up is maybe necessary.

The default stop script is allowing you to discriminate the stop operation
from a RUNNING system versus a BOOTING system by checking the MEMBER_STATE
variable.

>    steamboat ASE: mctahoe Agent notice:
>    /var/ase/sbin/lsm_dg_action: coldg: Disk group bench01_01_dg: No such
>    disk group is imported
>    ...voldg deport of disk group bench01_01_dg failed
>    ...voldisk: Device rz130: Device is already offline

Thus, these messages show you that ASE is trying to deport the diskgroup, but
is is not imported, then it tried to place the disk offline, but it is already
offline. These operations are part of the SERVICE STOP operation as explained.

Regards, Manu.

    Thanks,