Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

OS : WIn XP [COEM(DSP)/x64], SP2

RAID Controller : Adaptec ASR-2045

RAID Configuration : RAID 0

CPU : Intel Xeon Processor E5620 2.4 GHz

HDD : Hitachi 1 TB SATA3 7200RPM x5

I have 5 HDDS. My Computer configurations are - 1 HDD (C drive) has installed Windows OS. - 4 HDDs (D drive) has merged with RAID 0.

Sometimes we lost D drive. Have you ever experience this issue ? What is the root cause ? How to resolve this issue ?

Somebody said If there is some bad sector on any disk, then it fails and lost the HDD. Even though there is bad sector, I do not understand missing HDD.

Please, any comment would be very helpful.

share|improve this question
add comment

2 Answers

up vote 1 down vote accepted

Any time a HD drops out of a raid0 configuration, the whole of the raid0 configurations is failed.
Any drive that fails, you lose almost all of the data. only data that fits under the stripe size of what is left is recoverable. That means small files like text files.

If a drive goes to sleep, and does not wake up in a timely manner, a raid card will see it as lost, and the whole RAID0 array is then inaccessable. How the raid device and software reacts varies, with the more Pro raid cards (expecting a raid drive) having the biggest problems.

If a drive has a few minor errors, and the errors require that the drive move data from bad sectors to good sectors or reserved sectors, it also takes time for the drive to complete this task. Because the drive is not responding in a timely manner the raid card can deem it to be failing completly, even if you can wait till it finishes, reboot (as needed) and have the array operational again.

Any drive reguardless of sleep or error, or tler, could potentially not respond in general, or be very slow to respond when it is simply failing, in the many ways that a drive can fail. A drive that is simply failing, will usually get way worse over a short time.

Drives that are designed for Raid (general) can be set so they never allow sleep. Also they have a reduced time allowed for responce when correcting an internal error or moving a bad sector around. See TLER for more info http://en.wikipedia.org/wiki/TLER

Good raid cards are way more sensitive to this delay in responce time, and really cannot be used with normal desktop drives at all (mostly). OS Software raid and cheaper raid systems that are more software based are more lenient, and do not have a fit as quickly when a disk does not respond immediatly.

If your using a good hardware raid card, you really need totally raid compatable drives to go with it. Even then any of the drives could have a problem, and knock a drive out leaving you with 3/4 of each of the files, which is None of it :-)

If a single drive in the array is having problems, you could check that drive with SMART testing and see if it is gaining a large ammount of bad sectors. IMO(that is my opinion) when you assemble a raid system the drives should match in age, type, speed, quality.
If I had the problem you are having and the drives are old and have been spun up for long, I would replace all 4 :-( . I also would go with less risk and make 2 teams of 2 disks, not 4. Always have backups that are not raid based, and do not require the raid hardware itself.

share|improve this answer
    
RAID 0 configuration, it happens usually HDD missing. Am I correct understand ? I'm going to make RAID-5 or RAID-10 configuration. However, I need also high speed writing. Which one is the fast configuration ? Let me know other configurations. If I use CCTL utility, then it never miss the HDDs ? –  Changju.rhee Jul 31 '13 at 5:28
    
If it is for a desktop system, then backups can be better than mirroring. why? because mirroring the same data will also mirror a fail. Mirroring of data is excellent for servers where a hardware failure could occur and things still continue. raid5 is a lot of processing and rebuilds can take a lot of time, it uses less disks to achieve a redundancy through parity not Raw data. Raid 1&0 is raid0 with full mirroring. All of that is beyond my skill set, because I am not running a server. Making a good backup when everything is Good is my plan. –  Psycogeek Jul 31 '13 at 5:54
    
Thanks a lot for your answer. –  Changju.rhee Jul 31 '13 at 5:57
    
Take for example your situation 2 sets of 2T raid0 arrays (instead of 1 4T) If you use only 2T, you could manually sync to the other array very quickly, only after knowing that your system is stable. Sync software will also indicate what has changed. That would be fast, redundant, and if the user is very aware more safe. –  Psycogeek Jul 31 '13 at 6:02
    
2 Sets of 2T raid0 array mean? 2T is D drive with raid0. Other 2T is E drive with raid0. Am I correct understand? –  Changju.rhee Aug 1 '13 at 3:50
show 4 more comments

This is the downside of RAID 0. Any failure on any of the physical disks is a failure of the logical disk. RAID 0 has no redundancy whatsoever and should only be used for data that is either unimportant or regularly backed up.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.