X25-M G2 160GB / Fake commited IO after a power outage !

[idata]

With its write cache enabled, a X18-M X25-M or X25-E is expected to WRITES all io that is flushed...but it looks like it does NOT.

...and disabling the write cache looks like rendering so bad SSD write performance that one should stay in the HDD market...

Anyone has some INTEL official posts about this "pull the plug" nightmare for their SSD ?

NB: Please read http://opensolaris.org/jive/thread.jspa?threadID=121424 where they had some reproducible "pull the plug" test using basic io scenario (dbms used to write increments and display current value on the terminal, power outage, reboot, check the difference between the dbms and shown numbers)

[idata]

The topic "http://www.anandtech.com/show/3965/intels-3rd-generation-x25m-ssd-specs-revealed Intel's 3rd Generation X25-M SSD Specs Revealed" is not on top of Anandtech's storage folders any more

http://www.anandtech.com/show/3965/intels-3rd-generation-x25m-ssd-specs-revealed http://www.anandtech.com/show/3965/intels-3rd-generation-x25m-ssd-specs-revealed

[idata]

Windows itself may cache writes. It is also well known that Intel G2 drives do not guarantee writes just before a power loss, that is why they are labeled 'M', for Mainstream use.

[idata]

Apps that requires to bypass Windows writes can do it easily.

MSSQL has been doing that for years.

...but knowing you can't rely on a COMMITed io from a G2 drive deserves a rogue penalty.