9 TB Raid0 File Server

writethrough/writeback is for perc cards only (well, that i am aware of anyway)
Considering the performance achieved from the card for the price i think it is a worthy investment when dealing with raid

Copying a 3gb file from one of my other drives took under 20 seconds, averaging 140mb/s
Copying a 3gb file on the drives themselves took under 10 seconds averaging 250mb/s

 

RAID 0 is brilliant for people who don't give a shit about their data..

great for scratch disk, windows temp files and internet cache ..


anything else, you're a bloody idiot.

 

Redundancy =/= backup, anything which is raid is not a backup solution, it never will be, it's fault tollerence

Battery Backup Unit, it's designed to keep the data in the cache ram in the event of a power outage and when the system next comes online it will proceed to write the data, whereas if you didnt have a BBU, you could potentially break your raid or throw some serious file errors

 

You're right that redundancy and backup are different things. For example one big RAID5 is not a backup, but rather only redundancy. But two RAID5s where one is a delayed copy of the other, is nothing different than a backup. The fact that the backup is RAIDed doesn't change anything about its purpose.

In fact, most (serious) backups are RAIDed. In my case that means a second fileserver which acts as nighly backup for the main fileserver. Should the main fileserver be unavailable for whatever reason, then i can quickly switch to my backup server and continue on work. That is one benefit of having backups, amongst many, and it doesn't require you to leave RAID alone.

Many storage devices are also using RAID internally, including virtually all SSDs. That doesn't make them a poor backup medium.

A backup is a means of storage which is redundant in itself; you store things twice on different media. Whether that media is a RAID array or not doesn't change anything about the fact that one is a backup of the other.

A RAID array can not be a backup volume? I think you would want to re-evaluate that

 

Yeah exactly.

My only concern of a setup like two raid 5's or 6's in a single system is many failures can ocour, most notable being damaging power fluctations caused by power surge or power supply malfunction. in the end, in theory it is a backup, but nothing compared to offsite storage in either Bluray or LTO-4 Tape format

 

BluRay? Optical media storage has proven very unreliable in my case. If i had to choose, i rather store my valuable data on a 10-disk RAID0 than on a single blu ray disc. If there's something i do not trust, it is floppy drives and optical media.

You also assumed the two arrays are in the same physical system; i see no reason why that has to be and all the reason why you would want to separate the two, having their own system and power supply. That still leaves risks like Fire, Flooding, Earthquakes, Saddam's hidden missle-barrage, North Korea nuclear ballistic missles; yes if the world instantly ceases to exist if your lost your data, then you would want plenty of security mechanisms in place.

Seriously, risk management is often viewed without relation to the importance of the data. Having 20 off-site backups of the same data might make sense for something like a large commercial bank, but it doesn't make sense for someone who wants to store unimportant stuff; mostly illegally downloaded files. For that kind of data, implementing data security would be very different.

For most people, having a true backup would increase their data security considerably. Unfortunately, very few people have a true backup; especially if their data volume is quite large; like 9TB in this topic. Backing that up might require extra cash that the user is not ready to spend. In that case you have to make choices; which data is not important that won't have a backup and which data is very important that would want a backup and some additional protection as well.

Data security is different for each person. General rule of thumb is that you should be able to deal with one total HDD failure within 24 hours time, without causing headaches or a bad day, and recovery time within 24 hours. If your system is designed to cope with this, you should have a fun time without any headaches and assurance of your data security.

Also, using ZFS or other checksumming mechanisms is very wise, as without it you have no true measure about data corruption/integrity.

 

What drives are you running?

 

I have seen small companys that choose to use optical media, it isn't that unheard of albeit a bad method of doing so.

Of course i was talking about proper backing up, of critical data, not 10TB of linux isos which arn't that important

 

OP, what card are you using to get your 8 ports? Or does your m/b have all 8?