Severe performance issues are possible on Solaris for databases that use DISM. Although the most of the support notes have the same root cause (Bug 15794950), there is another problem affecting pre-12c databases that hasn't been well documented.
Checksum offloading causes bad (empty) packets in OVM for SPARC. (LDOMs).
V$SYS_TIME_MODEL, V$SYSSTAT and AWR reports show zero for CPU time in Solaris 11.1.
Veritas file fragmentation can affect Oracle database performance.
While investigating rsync/ssh performance, I discovered that our Solaris Clusters had just over half the expected throughput on their gigabit NICs. This was caused by the cluster/streams related module clhbsndr being automatically pushed to each public NIC via /etc/iu.ap . Symptoms: Round trip time via ping was higher than non-cluster servers on the same network … Continue reading Solaris Cluster vs Network Speed
Do you want to make scp and rsync go faster? LAN One easy method to squeeze a little more speed out of file transfer tools using ssh by trying a different encryption method. For example, we saved 15 seconds per gigabyte sent across our LAN on our SPARC Solaris servers, by using: rsync --rsh="ssh -c … Continue reading Encryption vs Transfer Speed
You may have read my advice not to use sparse tempfiles on Solaris UFS, but did you know that datafiles can be sparse too? Direct and concurrent I/O is very important for Oracle database performance, so there can be a dramatic impact when database operations switch to buffered and blocking I/O. (Eg RMAN section size … Continue reading Danger: Sparse Datafiles Ahead