Hi, we've finished our migration from 1.7 to 1.8 in our development sysplex and we've seen a strange behaviour with JES2 checkpoint.
Our configuration is CKPT1 allocated in CF and CKPT2 in DASD and this strange behaviour has been appeared in both CKPTs when all systems were migrated to 1.8 (correct values in a mixed env.) CKPT1 structure has increased (300%) its async req/sec and %delayed requests has increased from 5-10% to 15-20%. CKPT2 has increased its connect time from 5-7 to 12-13. We know that JES2 writes many blocks of data at once but it seems like JES2 control blocks have increased their length. APAR OA21560 explains that new code is introduced in 1.8: This is related to the new code introduced at 1.8 that uses XCF to signal other JES's telling them to wake up and get the checkpoint. This process is not known to 1.7 (or lower) members, and are therefore not getting a fair chance to access the checkpoint and execute jobs. Maybe this new code could increase req/sec in CKPT allocated in CF although we doubt that it increases (x3) the req/sec. We don't use the new message based protocol in CFRM and we've working with MASDEF,WLMPERF=OLD (red alert related with OA20680 and OA21212). Also, we've revised our CKPT1 size and it's correct. We've opened a PMR to analyze it but we'd like to know if you're having this effects. Thanks in advance. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
