|
From: Aaron J. <aja...@re...> - 2014-04-24 17:20:50
|
So I noticed a problem creating a sequence on my cluster - the logs indicated that there was a gtm failure. The GTM indicated that the sequence existed. So I stopped the GTM but left the two coordinators (and proxies) running. I then wiped out everything exception the configuration file and restarted. Soon after that, I noticed that the schemas I had created were no longer available through the coordinator - through attempts to create them said they existed. The best I can guess is that my reset caused the transaction #s to get munged and the schemas were not visible at the mvcc level I was currently querying. Long story short, I did this one to myself - however, in a real scenario, how would you go about resetting the environment if the GTM suffered a failure and needed to be rebuilt? |