FlexRaid: Transparent RAID (tRAID) vs RAID over File System (RAID-F)

[xnappo 1611]

 

To be honest, I value way more the pooling than the fault tolerance in RAID-F (and even the pooling has issues that needs to be resolved)

If one of my drives would fail right now, I'm not even sure if I would risk a recovery vs accept that my data is gone (I use "autofolderpriority" so it doesn't split across drives)

I have a batch file that list all drive content in txt just in case I need to redownload/rerip lost data.

 

 

Now you are confusing me again.  What do you mean 'risk'?  There isn't any risk in trying a recovery - it won't hurt the data on the good drives, and will only recover data that it can on the newly replaced drive.

 

xnappo

[sfnetwork 514]

Now you are confusing me again.  What do you mean 'risk'?  There isn't any risk in trying a recovery - it won't hurt the data on the good drives, and will only recover data that it can on the newly replaced drive.

 

xnappo

Yes it will, that the confirmation I asked Brahim in the past.

[ebr 16169]

Well I wouldn't go that far.  MOST of the drive B data will match MOST of the parity 'PAR' files, and will recover.  It isn't like one big PAR set, it is a bunch of little ones.

 

xnappo

 

Oh I understand how the parity works.  It is one bit for each combination of bits on all the other drives.  But that means that, if two of those other bits are changed from what was saved, that bit is going to have a 50% chance of being recovered incorrectly.  The thing is, there is no way to know where this will happen and how disastrous it might be.

 

However, it does seem that whatever damage there is would be limited to the failed drive's contents...

[sfnetwork 514]

you know what, I'm starting to be confused myself now, lol

This had been a while when I was asking this to Brahim, I'll try to find the tread in question to confirm...

[xnappo 1611]

ebr - yes that is true.  The data loss could be something you care about, or could be something you don't care about.  However FlexRaid 'knows' when it can't recover.  It is easy to see how bad it is, just run an update, change a few files, and run a verify.  It will tell you what you would lose based on that amount of change.

 

sfnetwork - naw - the data on the good drives is not touched during recovery. 

 

xnappo

Edited December 9, 2013 by xnappo

[sfnetwork 514]

Here is the original tread 

I am so sorry about the other drive thing, I was wrong. (lol trying to do several things at the same time here and my brain overheated  )

 

http://forum.flexraid.com/index.php/topic,1335.0.html

 

So mod put all data at risk between updates (but only of the failed drive recovered)

Edited December 9, 2013 by sfnetwork

[sfnetwork 514]

Also very important to enable the recycle bin to not worry about deleted file creating corruptions on top of this...

Edited December 9, 2013 by sfnetwork

[xnappo 1611]

Yep - that is right.  The issue is that with this statement:

- If the edited data is on other drives, recovery will be compromised and the restored drive will sustain random corruptions up to the size of the data edited on the other drives

 

You can be really hosed if you have 500mb of changed metadata - because that 500mb can be ANYWHERE, and of course just 1 byte of corrupt data can destroy a 8gb movie.

 

Keep in mind though that NEW data is just fine.

 

So.. How do we stop MB3 from using the RAID for metadata again?

 

xnappo

Edited December 9, 2013 by xnappo

[sfnetwork 514]

Yep - that is right.  The issue is that with this statement:

- If the edited data is on other drives, recovery will be compromised and the restored drive will sustain random corruptions up to the size of the data edited on the other drives

 

You can be really hosed if you have 500mb of changed metadata - because that 500mb can be ANYWHERE, and of course just 1 byte of corrupt data can destroy a 8gb movie.

 

So.. How do we stop MB3 from using the RAID for metadata again?

 

xnappo

By unchecking this I presume (never tried it)

[xnappo 1611]

By unchecking this I presume (never tried it)

Now that there is an awesome MB3 plugin for XBMC, I think I need to flip that switch

 

xnappo

Edited December 9, 2013 by xnappo

[sfnetwork 514]

Now that there is an awesome MB3 plugin for XBMC, I think I need to flip that switch

 

xnappo

I don't know how it affects the NFO files creation when this is off... I would guess it doesn't affect the XBMC saver.. I don't know,

@@Luke may confirm this....

[Luke 42077]

yea the plugin will save to nfo's to folders regardless, but you still need that for images

[xnappo 1611]

Yeah, don't need the nfos anymore though.

 

xnappo

[sfnetwork 514]

Yeah, don't need the nfos anymore though.

 

xnappo

If you don't actually use XBMC (I mean just XBMC) I guess not 

But it cool that it does it (in case you need it in future... My point of view...

Edited December 9, 2013 by sfnetwork

[xnappo 1611]

I do use XBMC - with my MB3 addon nfos are not needed.

 

xnappo

[sfnetwork 514]

I do use XBMC - with my MB3 addon nfos are not needed.

 

xnappo

Yes, thanks to you!

I need to get back testing your addon (so much cool stuff to test and not enough time)

[Airbender 128]

Check their wiki, it's pretty well described.

 

http://wiki.flexraid.com/category/traid/

Hello Sfnetwork

 

I been looking in their wiki and fourm my confusion is which software do i need if i want to go with T-Raid

The Host Package or Client package or both

 

Example: in MB3 i know i need server software in server and client goes to client pc but with T Raid it is not clear which one to go with and how to upgrade from F Raid to T Raid

 

Thanks

[sfnetwork 514]

From what I understood, both on the server (client one once, don't have to install everywhere...)

 

Please let us know if you test the tRAID because it's a different pooling system and it might work around some issues that RAID-F pooling has (access denied in logs, etc...)

[Airbender 128]

From what I understood, both on the server (client one once, don't have to install everywhere...)

 

Please let us know if you test the tRAID because it's a different pooling system and it might work around some issues that RAID-F pooling has (access denied in logs, etc...)

Hi

 

I am thinking about it as i did not pay yet for F Raid so i have options here but my problem right now is that Parity build takes time and i cant find any docs on Upgrade from Raid F to Raid t.

 

See i spent hours trying to figure out how to upgrade and not rebuild the Parity drive config and i cant find any data and guys at FlexRAID fourm are not very response and the mod guy he heat me big time because i ask to many questions or provide logs or screenshots

 

mybe i will ask in avs fourm and see if someone can help

 

i will update you guys if i figure it out

 

Thanks

[sfnetwork 514]

If you want to keep your parity drive (without rebuilding it), that won't be possible, you'll have to rebuild.

BUT, my question (which I never really had a solid confirmation yet), can you try the pooling without building a parity drive in tRAID?

[Airbender 128]

If you want to keep your parity drive (without rebuilding it), that won't be possible, you'll have to rebuild.

BUT, my question (which I never really had a solid confirmation yet), can you try the pooling without building a parity drive in tRAID?

Hi i am sure you asked or research this but isnt the Just pooling of all DRU will be under the advance tab of the config ? i believe the developer needs help with documentations on tRaid 

[sfnetwork 514]

Hi i am sure you asked or research this but isnt the Just pooling of all DRU will be under the advance tab of the config ? i believe the developer needs help with documentations on tRaid 

I don,t know, maybe you can start it of without a PPU but like everything, I would have to test it...

[Airbender 128]

I have my htpc pc that I am not using now as I am setting up my server I might be able to test and report back

 

 

Sent from my iPhone using Tapatalk

[Shaun 93]

1. TRAID Write performance

 

I'm using TRAID now.  I've had a lot of problem with write speed. I've been working with Brahim (the developer of FlexRaid) to improve the situation.

TRAID relies heavily on Random IO and I've built my NAS with WD 3TB Reds which have good streaming IO performance but suck at Random IO. I have improved the situation by 

 

•  moving from 2 parity disks to 1
•  Moving my parity disk from a WD 3TB Red to a WD 3TB Black (much better random IO)

 

I'm still getting less that 50MB, but I am sure that Brahim with improve the solution, because he is not going to get anywhere with this product with Write I/O that is half the speed of storage spaces

 

2. TRAID V FRAID

 

TRAID solves the potential data corruption issues that FRaid has because TRAID is a kernel driver and calculates RAID in real-time, there is no issues with out of date parity for edits.

 

3. Drive pooling.

 

Who needs it. Use MediaBorwser collections to do you pooling. I have 4 disks that make up movies, each of which is added to my movie collection.

Its not really that hard to keep track of when you next disk is filling up, add a new disk, add it to the appropriate collection

 

 

Summary

 

• I have 10+ 3TB disks
• I will have a least 1 total disk failure in the next few years
• I want a recovery solution the works without potential corruption
• I don't want to potentially loose my whole collection for a single disk failure (RAID 5/6 rebuild with a disk failure on rebuild)
• I don't want 10+ 3.5W disks spinning 24 hours a day (almost 1KW per day in power, 15c, 50 bucks a year just for disks + plus wear and tear)

 

RAIDT fits the above criteria. If I have to sacrifice write speed to get it, so be it. 

 

shaun

[sfnetwork 514]

Thanks for this, I appreciate it.

Writing performance isn't an issue since my concern is more the streaming from the clients (and metadata, images and renaming) so I shouldn't hit those limits.

 

But pooling, I'm sorry, I went down this road in the past (no pooling) and for me, ABSOLUTELY A MUST. (and also Media Browser isn't the only app there in my environment, XBMC with SQL, PMS, etc...)

I will never go back dealing with the drives individually (a lot of automation like MB, CP, SB, MetaBrowser, etc...).

 

I really wanted to have feedback on tRAID pooling as it's very different and as Brahim mentioned to me, will be more supported/compatible with monitoring applications that scans it it (Media Browser, MetaBrowser, Plex, etc...)

 

Plus, tRAID logically, "uses" all HD faster than RAID-F on long term. (based on my pool setting in RAID-F.

 

BUT, all that been said, you still made me want to jump in and try myself...

Edited December 12, 2013 by sfnetwork