I've got all the tuning things enabled.... I suspect the performance report does not match reality. I expect most people start z/OSMF once per IPL. If so, the implemented shared classes give no benefit, because it is the initial start up - which creates the shared in memory cache that costs CPU. The second start of z/OSMF benefits from the shared cache. That's why I suggested saving and restoring the cache across IPLs.
Any performance measurement/comparisons needs to be done from a just IPLED system. With the first start of z/OSMF, all of the (jar) files need to be read from disk into ZFS cache. The restart of z/OSMF will benefit from the files being in ZFS cache, and so the second time OSMF will start faster, and use less CPU. I'm hoping to look into this... but I cannot get z/OSMF started with an NISTECC certificate (problems loading keyring) - it only seems to work with an RSA certificate. Is anyone using a non RSA certificate? Colin On Wed, 15 Jan 2025 at 08:45, Sebastian Welton < 000007116c5c4437-dmarc-requ...@listserv.ua.edu> wrote: > KCINDEX is one way to go: > > https://www.ibm.com/support/pages/apar/PH06678 > > > https://community.ibm.com/community/user/ibmz-and-linuxone/blogs/andrii-vasylchenko1/2019/05/07/improve-zosmf-startup-performance-by-disabling-kc > > Another good guide: > > > https://www.ibm.com/support/pages/system/files/inline-files/zosmf_v2r4_resource_requirement.pdf > > Sebastian. > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
