Re: SE 7.12 Lockup?

In article <5davgg$495@cssun.mathcs.emory.edu>, Joseph Cullipher <joseph@cannonexpress.com> writes >John, >I haved had this problem since...since...well it's been a long time. >It first appeared with SE 6.00UE1 on SCO 3.2.4.2 after about 2 months >with tech support trying to figure it out we upgraded to 7.10UC1 >The problem is still occurring and tech support still has the case open. I doubt the engine is 'locking up'. kill the frontend process i.e. .4ge (I normally use kill -11 - SEGV since anything should handle what might happen if it seg. faulted and few processes ignore it). They wait until the sqlexec finishes what it is doing and tries to talk to the 4ge. If you kill it before then it may be trying to rollback a transaction. If you kill it whilst it is rolling back you WILL get a corrupt database. >They have told me that I have been the only one to ever report this but >I have seen some similar problems listed in this forum by others. >At times the engine may lockup 6 or 7 times a week and then go 2 to 3 >weeks without a problem. We have been looking into TCP as a possible >cause but cannot find anything definative. If it happens to you again, >you should try a kill -15 on the sqlexec process. Make sure no user is >trying to connect or this will not kill the process(lots of experience >with this). If you need any more info or if you find the solution to >the problem let me know. > -- David Williams