C
c3rberus
Guest
In our Q&A we tried to upgrade from 11.6.3 to 11.7 (and 11.7.3 - the updated re-release) and we hit a wall. Our .NET application uses state-free AppServers. Couple of our scheduled async tasks started failing with protrace: uttraceback : 0x0000001c uttrace_withsigid : 0x00000150 utcoreEx : 0x00000130 drexit : 0x00000484 drSigFatal : 0x000000a8 fdGetTblPartLock : 0x0000010c screadlock@AF51_24 : 0x00000194 scidxb : 0x00000124 proidxcache : 0x00000120 fdidxcache@AF202_40 : 0x00000044 scfild : 0x0000015c IPRA.$rnreloc : 0x00001e08 rnproc_entry : 0x00001414 rninterpret : 0x00000064 IPRA.$rnudfunc_local : 0x000002d4 IPRA.$rnudfunc_Exec : 0x00000160 rnudfuncOO_Body : 0x0000022c rnudfunc_run : 0x00000170 rnudfuncOOMethod : 0x00000158 fmoo4glRunMethod : 0x0000010c fmoo4glGetAttr : 0x0000016c IPRA.$rnwdatrEval : 0x000000b4 rnwdatrx : 0x00000034 fmEWDAX : 0x00000044 fmeval : 0x00000330 rnexpstmt : 0x00000230 rnexec_entry : 0x00000278 rninterpret : 0x00000064 cr_run_loaded : 0x00000028 IPRA.$execProc : 0x000005f8 IPRA.$execCall : 0x00000034 IPRA.$processRequest : 0x00000458 ReaderClose : 0x00000284 closeWrite : 0x0000001c open4GLWriteLast : 0x00000098 IPRA.$ub_pushRq : 0x000004dc ub_processRequest : 0x00000200 csd_dispatch_message : 0x00000530 IPRA.$do_serve_mainline : 0x00000308 IPRA.$do_serve : 0x00000038 classicapsv_main : 0x000000d0 main : 0x00000070 We also saw: uttraceback : 0x0000001c uttrace_withsigid : 0x00000150 utcoreEx : 0x00000130 drexit : 0x00000484 drSigFatal : 0x000000a8 fdseqval : 0x0000025c fmENXTVAL : 0x0000004c fmeval : 0x00000330 rnasgeasy : 0x00000268 rnexec_entry : 0x00000278 rninterpret : 0x00000064 cr_run_loaded : 0x00000028 IPRA.$execProc : 0x000005f8 IPRA.$execCall : 0x00000034 IPRA.$processRequest : 0x00000458 ReaderClose : 0x00000284 closeWrite : 0x0000001c open4GLWriteLast : 0x00000098 IPRA.$ub_pushRq : 0x000004dc ub_processRequest : 0x00000200 csd_dispatch_message : 0x00000530 IPRA.$do_serve_mainline : 0x00000308 IPRA.$do_serve : 0x00000038 classicapsv_main : 0x000000d0 main : 0x00000070 We are able to re-produce the error by re-running the programs. I have a case open with TS. If we switch the Q&A environment back to 11.6.3 it works without issues. Running the same program directly (using mpro) instead of going through AppServer calls, seems to work 99% of the time. Adding some debug message statements before and after the procedure to try and figure out the records it is stuck on seems to cause the procedure to run successfully via AppServer. Anyone run into this? Any ideas? I would like to get over to 11.7.3 to make use of the new logging parameters but we've hit a wall. Our multi-tenant databases are running on AIX 7.1. Seems like a odd lock timing bug given the procedure success rate increases by simply adding debugging message statements.
Continue reading...
Continue reading...