磁盘IO不行.可以先分配空间.
http://blog.itpub.net/267265/viewspace-1107524/
google 找到如下连接:
http://dbasolved.com/2014/03/05/ ... t-enq-hwcontention/
http://www.oracledatabase12g.com ... BA%89%E7%94%A8.html
Opening the SR has lead me to dig a bit and do some research on extent allocation and LOBs while waiting on Oracle's
answer. Prior to Oracle Database 11g (11.2.0.1), there is a known bug (6376915). This bug is related to high watermark
enqueue contention for ASSM LOB segments (Note ID: 6376915.8). This bug was actually fixed in release 11.2.0.1 but it
needs to be "turned on" in later releases. To turn the fix for this bug on in 11.2.0.2+, an event needs to be set in the
spfile.
SQL> ALTER SYSTEM SET EVENT="44951 TRACE NAME CONTEXT FOREVER, LEVEL <1 .. 1024>" scope=spfile;
By setting this event between 1 and 1024 will identify the number of chunks to be cleaned up each time a reclamation
operation is performed. In turn this reduces the number of requests against the high watermark enqueue.
Once the event has been set in the spfile, the database instance needs to be rebooted to make the change effective.
--另外你检查看看交换空间是否充足,sun 是看 /tmp目录.