frozen recovery database

[Kevin Scantlan]

Had a situation where the recover db for a site was “frozen”.  What I mean by that is that no process including hcidbdump from the command line was getting a response.  No errors, but no response.  We thought we needed to bring the entire site down.  Started with a process.  Ran hcienginestop and it was not killing the process until it did a SIGKILL.  I was bringing down other processes with other sessions.  When the first process that had a SIGKILL done on it finally went down, then no other processes had to.  After bringing down 4 processes, noticed that the recovery db was no longer hung.  So we brought the processes back up and all is right with the world.  What would have caused this?  Nothing showed in the process logs.

At the same time (but not knowing it), noticed a process was down.  THE IDE monitor and hciprocstatus both agreed.  Process log showed nothing.  hciconnstatus showed all threads in that process were UP as did the IDE monitor.  Was able to bring up the process without a problem.  What would have caused the hciprocstatus and hciconnstatus to disagree?

Topic