What has happened since the prototype port of OpenSolaris to the IBM mainframe under z/VM ("sirius") that occurred several months ago? Jeffrey Savit\'s blog reports on the developments since then, and reveals what's still in the works.

What's New and Better

There is now a man and prstat command, a hostid command, and the problems that made getconf -a and 64-bit ls fail have been resolved, the blog reports, adding that the bug preventing a change of password has also been resolved.

Further changes involve correction of the error that resulted when users attempted to create a zone with OpenSolaris. The blog notes as well that zonecfg now apparently works, though zone installation fails with a complaint that a script related to Solaris Live Upgrade (LU) is missing, saying /usr/lib/lu/lucreatezone: not found. Clearly, the blog continues, there is an issue about implementing Solaris Containers in the way that is compatible with OpenSolaris boot and package system - both of which have yet to be ported.

Another necessity that needs to be addressed, the author contends, is that better use must be made of the ability of OpenSolaris to boot off ZFS rather than the Unix File System (as Solaris does).

Other missing pieces include DTrace and a package/patch management system, though the blog notes that the Image Packaging System is now going to be ported over. If you issue the format command, it still says "No disks found" (even though there are definitely disks available and mounted). The kstat and prtconf -vp commands don't work.

The author also notes the unpleasant experience of using commands like pstack or pfiles to diagnose what a program is up to only to discover it can die or kill the target process, as that set of tools isn't quite working yet.

What's Now in the Pipeline

The author found that setuid fixed the problem encountered in ping not working from a non-root userid. Still to be resolved is the issue with ping or ftp dumping core when DNS name resolution is used. While turning off DNS helped with this, the author considers the problem still requires attention.

The blog also notes the recursive crashes that occurred in the Service Management Facility (SMF) daemons, which the author suggests might be addressed by reloading the test system from the disk restore image (you install this by doing an image restore of a disk volume, and manually configuring network identity).

A final consideration involves the fact that sirius currently runs only under VM in order to simplify the porting effort. As a result, it is not possible to run sirius on commodity, developer personal computers. This is under study, the author adds, and should be resolved soon.

One last note involves the possibility of running OpenSolaris in a Hercules virtual mainframe, running inside a Solaris Container underneath OpenSolaris or Solaris 10, running inside a VirtualBox or VMware Virtual Machine. Perhaps it will become possible to bring up Hercules underneath sirius, for an arbitrary number of recursive levels of virtualization.

More Information

Jeffrey Savit\'s blog

OpenSolaris Project: Systemz

Get started with OpenSolaris [...read more...]

fullsource

From the latest issue:

Recent Blog Entries as of March 22, 2010, 10:00 am
Oracle Works to Assuage the Anxieties of Its Customers SysAdmin News Bites Oracle Enterprise Manager Ops Center Released Suits and Their Countersuits A More Direct Channel Model Planned for Oracle-Sun	The Importance of Information Storage and Management Sun POJO Service Engine User's Guide Sun Storage 7000 2010.Q1 Quality and Features PSA Deploys GlassFish in the Manufacture of Peugeot, Citroën Vehicles Running VirtualBox VMs as Services in Windows
Blog RSS Feed