Informix - Problem description
Problem IT20579 | Status: Closed |
ONMODE -K HANGING, THEN ONCLEAN -K AND RESTART WOULD LEAD TO SAME INSTANCE RUNNING TWICE | |
product: | |
INFORMIX SERVER / 5725A3900 / C10 - IDS 12.10 | |
Problem description: | |
In a situation where onmode -k is unable to bring down the server, then onclean -k is used in addition, onclean would not kill the still running oninit processes, but only 'clean up' shared memory and semaphor resources. As the old oninit processes still are attached to their shared memory, those shmem segment would only be flagged as deleted, but could not be removed. As a consequence nothing will prevent the instance from being started a second time, so two sets of oninits would now be running for this instance, and still the old instance would have the tcp ports open (the new one would run on 'cannot bind a name to the port' listener-thread assertions). So data modifications and physical/logical log traffic including checkpoints might be coming from two independent instances to same chunks. In addition twice the amount of shared memory would be allocated, potentially causing more trouble for the OS. | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * Users killing hung instances with onclean -k * **************************************************************** * PROBLEM DESCRIPTION: * * See Error Description * **************************************************************** * RECOMMENDATION: * * Update to IBM Informix Server 12.10.xC9 * **************************************************************** | |
Local Fix: | |
Kill server manually (kill -9 on first CPU VP) rather than or in addition to using onclean -k when onmode is unable to bring down the server. | |
Solution | |
Workaround | |
not known / see Local fix | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 12.05.2017 10.07.2017 10.07.2017 |
Problem solved at the following versions (IBM BugInfos) | |
12.10.xC9 | |
Problem solved according to the fixlist(s) of the following version(s) |