Frequent HDD Failures - Proliant DL380 G5

[arazas]

Hi,

We have 6 HP DL 380 G5 Servers. We have P400 RAID CONT with differnt config. Some servers RAID 1 and some are RAID 5.

This setup was installed in 2008-2009.

Now from last 6 months, We have recorded very freequent HDD failures. Every month one or two HDD are going off. showing amber light and failure logs in system events as well as in HP Insight management tools and RAID controler showing broken arrays.

Two times new HDD were also gone which were replaced recently.

Its a big trouble and huge overhead expense for company, I need to submit a true factbase report to management why it is hapening and how can we stop it.

Somebody from HP officials, please assist me with this problem.

P.S. this thread has been moevd from Servers > General to ProLiant Servers (ML,DL,SL) - HP Forums moderator

[Jan Soska]

It is hard to comment your problem.

In our enviroment, we have very low disk fail ratio. For example storage eva6000 with 96 FC disk only one (1!) failed in 6,5 year period. In ithrer hand - we have 8 fata 1TB disks in the same storage and 6 of them failed in 4 years. To be even more complicated, we have another MD3000i storage with 15 15k SAS disks and 7 of them failed in 4,5 years.

as you see, there is nothing to be said in general. From my personal point of view, I've stopped to believe in sata/fata disk, and it seems disk in higher storage classes (comparing midrange EVA and entrylevel md3000i) have better durability, maybe due carefull testing?.

I am quite curious if this continues in modern midrange arrays where FC disks are forgotten and SAS prevails.

Anyway, enterprise enviroment is all about support. If you pay support, you have suppport -  and your disk are replaced..

Jan

[krishblr]

There is  firmware and driver which needs to be updated as and when released by HP to keep the HDD firmware updated to latest to aviod failures. (some percentage of failure can be avoided with regular updates.)

Physical failure, due to power and enviornment cannot be avioded

[Med-H]

Hi,

Try to update the server with the last frimware to aviod failures :

http://h18004.www1.hp.com/products/servers/service_packs/en/index.html

thanks