Control the size of the ZFS ARC cache dynamically

Last updated on: 25th Dec 2014

Solaris 11.2 deprecates the zfs_arc_max kernel parameter in favor of user_reserve_hint_pct and that’s cool.

ZFS has a very smart cache, the so called ARC (Adaptive replacement cache). In general the ARC consumes as much memory as it is available, it also takes care that it frees up memory if other applications need more.

In theory, this works very good, ZFS just uses available memory to speed up slow disk I/O. But it also has some side effects, if the ARC consumed almost all unused memory. Applications which request more memory need to wait, until the ARC frees up memory. For example, if you restart a big database, the startup is maybe significantly delayed, because the ARC could have used the free memory from the database shutdown in the meantime already. Additionally this database would likely request large memory pages, if ARC uses just some free segments, the memory gets easily fragmented.

Read More

zpool-zones-vio.d - ZFS statistics per zpool and per zone

Another script, which provides a missing feature to Oracle Solaris 10/11: ZFS statistics per zpool and per zone.

The ZFS in Oracle Solaris 10 and 11 is still not completely ready for a single pool setup, for some applications it makes sense to use multiple pools. In general, if one workload stresses the single zpool too much, the performance of the other workloads can be degraded. Therefore it’s still best practice to use multiple pools, for example the “Configuring ZFS for an Oracle Database” white paper suggests to use at least an own pool for the redo-logs.

Read More

oracle-pwrite-zfs-latency.d: Detect slow sync writes on ZFS

The first DTrace integration into the Oracle 12c database was finally reverse engineered in great detail by Andrey Nikolaev. Finally we have a good documentation how the V$KERNEL_IO_OUTLIER view really works.

The standalone kernel_io_oulier.d DTrace script works great even on Solaris 10 and pre-12c Oracle databases, but only on RAW-devices and ASM. Oracle pushes ASM because of good arguments, but sometimes one likes to use a filesystem.

We have some databases on ZFS and so I was jealous that the script does not work with ZFS. Therefore I tried to make a similar script for ZFS.

Read More