Re: Online Engine Aborting -- Help!

In article <1994Mar1.162948.4900@lmpsbbs.comm.mot.com> ennis@ssd.comm.mot.com (Bill Ennis) writes: >We are experiencing a problem with the Online engine shuuting itself down. >We see the following in the online.log: > >4:30:17 Lock table overflow - user id 110, process id 23526 >04:30:17 Lock table overflow - user id 110, process id 23526 >04:30:22 Lock table overflow - user id 110, process id 23526 >04:30:22 Lock table overflow - user id 110, process id 23526 >04:30:27 Lock table overflow - user id 110, process id 23526 >04:30:27 Lock table overflow - user id 110, process id 23526 >04:30:32 Lock table overflow - user id 110, process id 23526 >04:30:32 Lock table overflow - user id 110, process id 23526 >04:30:37 Lock table overflow - user id 110, process id 23526 >04:30:37 Lock table overflow - user id 110, process id 23526 >04:30:56 Process Aborted Abnormally: pid=24710 user=1010 us=4040275c flags=1 >04:30:56 Process Aborted Abnormally (latch): pid=24710 user=1010 flags=1 >04:30:56 INFORMIX-OnLine entering ABORT mode!!! >04:30:56 -- Online Aborting -- us=40401b8c, pid=16995, uid=109> >Tue Mar 1 04:40:59 1994 > >04:40:59 INFORMIX-OnLine Initialized -- Shared Memory Initialized >04:40:59 Physical Recovery Started >04:40:59 Checkpoint Completed >04:40:59 Physical Recovery Complete: 0 Pages Restored >04:40:59 Logical Recovery Complete: 0 Committed, 0 Rolled Back >04:40:59 Dropping temporary partition 1000006, recovering 8 pages. >04:41:01 Quiescent Mode >04:41:04 On-Line Mode >04:46:25 Checkpoint Completed >04:59:42 Checkpoint Completed >05:00:22 Checkpoint Completed> >The Lock table overflows actually started at 3:50 AM. Informix claims >that the process aborting abnormally was because a process was killed. >I don't buy this since users do not have Unix level access. > >Thanks -- Please replay via email, Bill, Edit your TBCONFIG file. Increase the LOCKS parameter greatly. Locks use very little memory. Who is user 110? Are they an enduser or an admin person using ISQL. Our lock overflows are most often caused by a person doing a global modify with SQL or forms and not explicitly locking the entire table. Informix was right (partially). The process (24710) that caused the abort probably died internally because the lock table was full. ----------------------------------------------------------------- Jim Burks jburks@promus.com The Promus Companies Memphis, Tennessee, USA