Re: Can't IPL because disk drives are full

On 11-Jun-2010 17:08, Albert York wrote:

I have a system at V5R4 which won't IPL because the disk drives
are full. I can do a manual IPL and can get to DST (which shows *
for disk usage) but if I try to continue with the IPL it tries to
do a dump and dies.

Is there a way to clean out the jobs logs and output queues from
DST? Or do a RCLSTG from the panel?

Once I get the system up I can do some cleanup so it will IPL
normally.

Any suggestions would be appreciated.

Perform a Manual IPL and two "change system" options should be presented; IIRC, if the signon screen is reached for OS IPL phase, then just after signon. Choose both by Y=change [or whatever the input requirements, to effect "Yes I want to change"], and one of those two options should offer the options for both "Clear all job queues" and "clear all spool files". Choose to do both /clear/ options, cross your fingers, then continue the IPL.

If those options are highly negative and best would be avoided if at all possible, then there are also some /tricks/ that can be effected under the other /change/ option if those menus can be reached; they should, if the other noted options are available on a manual IPL.

There is also an option to hard damage the database cross reference file dataspace(s) [QADBIFLD for example, which is typically a very large file on most systems], which will effect the deletion of the data, its member, and its file... with all the very negative side-effects. Note: hard /damage/ does not apply to object type x'19', thus do *not* ever damage the *FILE object itself, and the IPL review of the dbxref files is on the dataspaces [and IIRC the members].

Regards, Chuck