At SAP TechEd, I sometimes get ideas on useful time savers that I should be able to employ right away, not needing to wait until the next version, the next patch, the next year. Turns out I saw a feature I’d seen 2 years ago, and still don’t have it where I want it.
- ST06N/OS06N/OS07N – A new era for OS collectors
- Posted on Oct. 13, 2007 07:27 AM in Application Server, SAP TechEd
My blog post 2 years ago showed what was possible under the then-named Solution Manager 4.0. That was our non-Unicode system, and we had just begun to set up central performance history. Where are we now (are we home yet?)?
Tony de Thomasis showed the Australia Post remote OS performance data from Solution Manager in his TechEd session (ALM218). Don’t look for this in his slides, it was done live. That makes it a wee bit harder to duplicate.
I started with OS07 (“Remote Operating System Activity”), getting a pop-up for some of our sandbox, dev and test systems:
OS07 pop up
The initial drilldown into each of these yielded a promising screen (“Remote (SAPCCMSR…) / Operating System Monitor AIX”), with no data, unfortunately.
The status message was correct, yet incomplete and unhelpful (aren’t they all?)
Partner SAPCCMSR.HOST.99 not reached (reading table )
Not completely unhelpful, I suppose, as the hint was about “SAPCCMSR” rather than the “SAPOSCOL” daemon I would have suspected. Yes, once again, the daemons are not running on the satellite systems. Need to fix that.
Once we kicked off daemons on a couple machines, I went back the next day to look again. Success! A minor win, in the scheme of things, which we’ll take.
OS06 remote after
The status message reported seems to be rather minor. I can’t find anything missing, and suspect that the “system problems” may result from network drives on the host system and don’t truly mean that a complete system has failed. On the other hand, we need to report this through OSS/Service Marketplace and I’m not holding my breath for an easy answer (nothing found on my first searches – yes I looked).
That’s a snapshot, which as commenters to my prior blog pointed out, are hardly a match for OS-specific tools such as vmstat, top, topas, nmon, or [your favorite scanner here]. What about history and trending?
The “Detailed Analysis” menu choice of remote OS06 does show the remote system choices as “Current Data for Selected Server”, both Snapshot and “Previous 24 Hours”. These are highlight with the red bar to the right.
Remote data on an hourly basis is visible within Solution Manager, according to the title bar:
OS06 remote 24 hours
On the downside, the menu item highlighted in blue does not take me to the remote system data I expected to see, but shows the local Solution Manager system, again, per the title bar. Why doesn’t this work as it should?
OS06 “remote” 30 days
Another way to get to the same data is through OS06N. Running OS06 shows just the current environment, but the “N” will get you a tree view of systems that are remotely accessible. Two years ago, the screen showed a folder with “Application Server” as the organizational unit, but today with Solution Manager Enhancement Package 1, I see “Hosts” and “Systems” as the 2 top level folders. The former shows data on the frame level, not just the running partition, so I can’t wait to get this set up on the fully virtualized landscape we’re now building out.
The latter shows application server instances running in each node (e.g. “DEV”), yet they also include nodes that have not existed for years. I’ll need to find the purge button to clean out that basement.
And yes, there are communication errors on many of our hosts. Agents are probably not running, remote systems not trusted, or may have been set up with hard coded machine names instead of aliases.
All in all, a good start, and a vision of monitoring trends of the future.