Should read D:\oracle\base\oradiag_SYSTEM\diag\clients\user_SYSTEM\host_1017316664_80\trace, but limited by characters in the title.
I have a two node RAC cluster on Windows 2008 64 bit.
Oracle version (DB and CRS) is 22.214.171.124
When the first node is rebooted, the file system (D:) fills up and the only real directory I've found that is filling up is the one mentioned above. It wrote over 100,000 files there last time, all trace files with unable to connect to ASM instance.
I have to delete the files in here to get it to continue, however, then the D: drive appears to 'drop out'. The D drive is a local drive on the server and is where the oracle binaries are stored.
Running crs_stat -t did not return any output whatsoever, so suggests the crs was in the process of starting.
I left this overnight and in the morning, all services were running as they should, however, I have no way of finding exactly what time this resolved itself.
And a generic statement to everyone running RAC on Windows ... Why? Have you not paid attention to all of us that keep saying "Don't do it" because of the multitude of issues most important of which the glacial speed of failover?