Compact Flash Card Reliability

[Scubatooth]

for non technical types sorry for this


well raid 0 is worthless i will agree with you on that.

just a question why would you use a raid 3 array for anything because the DW my dad works on is a 7 TB array running on a sun spark 32 proc system with a EMC frame running a 0+1 setup and they dont have any problems with data loss at all thanks to the emc frame set up that is constantly monitoring for faults(and once the threshold has been broken it calls home for a tech) inaddition to this there is a vertas volume manager monitor everything.

my question is why have you choosen to use raid threebecause that requires a drive for the checksum and error correction process ( i can only think of the video editting a need for this doing a-b nonlinear editting) and then 5 is just way to slow in write time for a DW thats has 25,000 instances running at any given time.

? what industry do you set up these machines for?


but personally for my systems i run a raid 1 ish style set up in that i have a main system harddrive (WD raptor 10k ) that has the sys install with all of the programs, but then the next drive is another raptor that is used as a scratch disk (for graphic and photo) but this is only limited to around 3 gb on the drive the rest is used for a active ghosting of the drive (set on 3 hour timers) that will ghost the sys drive and keep a image on the drive and i keep 3 images worth so that if one is bad i havent lost more then about 6 hours work (not a killer but redoable) then i have a similar set up for data which is never on the system drive. i use ghost walker for ghosting setup

layout

C: System drive (10k WD Raptor)
D: Scratch Drive (this is for program and OS scratch and page file ~3gb worth, the rest is system back up) --(10k WD Raptor)

E,F,G optical devices DVD Rom reader, DVD Burner, Hot swap HDD cage

H: Onboard System data storage (WD 120 GB JB)
I: Backup (120 for sys 120 for external HDD for laptop)

remaining 10 GB is used as a back up scratch disk, but this may change a ramdisk soon once i get the software



FYI a TB is equal to 1000 gigabytes

 

[chinacat46]

One of the reasons to use RAID 3 or 5 over 0+1 is that it's cheaper. Lets look at a simple 4 disk RAID with each of the above. In RAID 3 and 5 three of disks are used for data and the other for parity data. Thus 3/4 of the total available volume is being used to store data. If one of the drive goes down you can still have availability of your data and can call home just like a 0+1 so the calling home thing is not an edge for either of them. On a 0+1 with four drives you basically use two drives for striped data and then mirror those so you only have data on 1/2 of the drives so you get more data on a RAID 3 or 5 system those lowering overall cost. RAID 3 systems are generally used for video Production and live streaming, Image Editing, Video Editing, Prepress Applications and any application requiring high throughput. RAID 5 systems are generally used for File and Application servers, Database servers, WWW, E-mail, and News servers, Intranet servers and overall is the most versatile RAID level. As for RAID 0+1 it is generally used for Imaging applications and as a general fileserver. All 3 have advantages and disadvantages. Since you seem to think 0+1 is the best lets look at some of its disadvantages. For 0+1 to store the same amount of data as the RAID 3 or 5 systems I've described(3 data disks 1 parity) you need 6 drives. Three for striping and the other 3 to mirror that. So you can see you are paying for 6 drives instead of 4 and RAID 5 and 0+1 have the same fault tolerance. With RAID 5 you loose a bit in writing performance for the cost advantage. RAID 3 has good performance on both read and write. RAID 0+1 has the highest performance. Scalability is another issue. If I add one drive to a RAID 3 or 5 system you must add two drives to a RAID 0+1 to store the same amount of data.Thus if I had a RAID 5 with 30 drives. I would have 29 drives for data and one for parity. The 0+1 would require 58 drives to store the equivalent amount of data. That is a lot of overhead for the same fault tolerance and most companies can live with a little less performance. On a database server or a www, email server most of the transactions are reads and thus the fact that writing is slower is not a big issue. As I said all 3 have advantages and disadvantages. But as I just illustrated the inherent cost of 0+1 is the main reason why Level 3 and 5 are much more popular then 0+1. Just because you know one DW that uses it doesn't make it the end all.

 

[Scubatooth]

china cat

im not disagreeing with you on the merits of the different raid levels as each has its own area which its application is ideal, i was just showing the example for which the company my dad works has set up there data warehouse and it works for them

old adage to each is there own.

Tooth

 

[chinacat46]

No worries this whole thing started though cause you didn't state the chip was emulating Level 0 and made a generic statement that it was like a RAID device where if one disk fails the whole system goes down. While this is true for level 0 it isn't from most RAID devices which is what I took issue with. I think we both agree that Level 0 isn't redundant so there isn't any issue. The whole purpose of RAID is redundancy which is why I don't consider Level 0 a true RAID. I understand that the XD is striped like level 0 so in that case you are right if one chip goes down it's useless. Forgive my bad grammer and spelling I'm an engineer. Engineers like to argue.

 

[Former FLL Now MSP Diver]

Now that I ordered my Oly 5050, I have read the sites that talk about CF cards and various speed ratings. That is not an issue for me but reliability is. I just read this article at http://www.kenrockwell.com/tech/cf.htm.
Interesting to say the least. So have any of you lost photos/had corrupt files on your CF card? The price for a 512M CF card ranges from $75 for a Viking at Amazon (with rebate) to up to $180 for Lexar and others. Please post your experiences with different card manufacturers like Transcend, Sandisk, Viking, Kingston, Lexar, etc.

TIA

Reeling the conversation back in from making RAID arrays out of memory cards...

I've had three different brands of cameras that have all used CF cards, and I've never had a corrupted card or have lost photos. A couple of the cards are now over 5 years old.

I did manage however to bugger up them myself. I couldn't figure out why the cards wouldn't work after I formatted them in the card reader attached to the computer. I thought I had given them a static charge and fried them. Then I read somewhere that cards formatted to FAT32 weren't compatible with the camera I was using. I took the cards I thought I had fried and formatted them in the camera, and now I have more cards than I know what to do with. :crush:

Marc

 

[Scubatooth]

Fll diver

that is one wired thing about digital cameras is that each manufacture has its own version of a format for the cards, which im most cases is far from the normal fat, and then anything above fat 16 isnt readable by the camera unless its one of the new high end DSLRs.

so the best reccomendation is to format the card only in your camera and not on your computer. also dont worry about having to many cards as that is a good thing because i have over 2 gb(1 1GB, 1 512, 1 256, 2 128 and will probably get more this summer since i will be getting a DSLR in august ) worth of CF cards and i have run into instances were that wasnt enough(shooting a Canon 1D at a high school football game last year in RAW+Jpeg mode, that will eat up cards in a hurry) so i keep plenty handy and the laptop to.

Tooth

 

[String]

Ive used Sandisc and Kingston cards of various sizes and never yet had a single problem.

Ditto Smartmedia.

 

[Big Blue One]

.. basically by removing it from the reader while it was still in a "writing" state ie. light flashing or orange.

I subsequently discovered this was rule #1 and after trying many reformatting, recovery etc. solutions discovered why - the card was completely unsalvageable.

Otherwise had absolutely no problems.

Just have to watch if you remove the catalog files from the card with pictures still stored on it, your camera will indicate it has apparently less available memory but you wont be able to call up the pictures to delete them. This will have to be done straight of the computer