B3 RAID status problem

Got problems with your B2 or B3? Share and get helped!
mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

B3 RAID status problem

Post by mcrsro » 20 Feb 2014, 08:47

Raid_failure_1.png
This is the screen showing the failure. Another screen shows that the main disk from the RAID (containing the system) is depicted in RED.
Raid_failure_1.png (647.45 KiB) Viewed 7996 times
Raid_failure_1.png
This is the screen showing the failure. Another screen shows that the main disk from the RAID (containing the system) is depicted in RED.
Raid_failure_1.png (647.45 KiB) Viewed 7996 times
Dear Gents,

First of all I want to thank you for B3 design which I and my family used intensively on our home network.
I bought my B3 on March 2011 in a configuration with an additional S1 unit and I configured to work in a RAID mirroring configuration.
It was proved to be very useful, and the design can be appreciated as being very good.

The server worked well till last month, when I noticed that the error shown in the attached picture.
Unfortunately, in another screen the main disk from the RAID (which has installed the system) is colored in red.
These symptoms show, in my opinion, that the disk located in the B3 box (a 2 TB WDC WD20EARS-00M disk, which is from the Green series) is defective.
This means that I have to backup everything, to replace the disks and to reinstall the B3 software, reconfigure the RAID and restore the saved data, because on the defective disk located on B3 is installed the operating system.

I consulted the disk data sheet and I found that this original disk has a warranty of only 2 years and, because my server is more than 2 years old, it seems that the best solution is to replace it. Unfortunately, this disk is an old model and is no longer available. Much more, I suspect that the second disk from the RAID might fail in the future, because it has left the warranty period. I want to avoid in the future any escalation of these type of events.

Same time I would wish to replace both disks with a WDC model WD2002FAEX that has 5 years warranty and is readily available. This is so-cold "Black Series", which is faster but is not a "Green" version. Its consumption is 10.7 W instead of 6 W (which is the consumption for the WDC WD20EARS-00M model).

From my viewpoint, I have no objection regarding the increase of total power consumption of the server ( instead of, let's say, 2x6 = 16 W I will have 2x10.7=21.4 W ).
The advantage would be the increased speed (I've seen on the excito forum tha some users had problem with IntelliPower feature integrated into the "Green" series) and, much important, the increased reliability (it must be so because WD offers 5 year warranty for the Black model).

I would wish to obtain from you an answer the following questions:

1. It is allowed to replace the 2 TB WD20EARS-00M ("Green Series") with the model 2 TB WD2002FAEX ("Black Series")? I would wish to tell me if the power supplies from B3 and S1 module may support the model WD2002FAEX.

2. Western Digital offers other model with 4 TB, having the same power consumption characteristics like the version with 2 TB. Is it possible to use this type? In this case, is it possible to configure the RAID as a 4 TB mirroring assembly?

3. What is the best succession of operations which can restore the RAID functionality? I mean, is it a better procedure to restore the data from the existing good disk from the RAID ( it is the HDD mounted into S1 module), avoiding to backup the data, but avoiding too the data loosing? Please describe me the most recommended succession.

Thank you in advance,

Cristian Moraru

PS. I purchased the B3 on 2011-03-08, my system containing the items below:
- B3 miniserver code 1.00B3W-2000
- S1 SATA disk for RAID 1.00BS-2000

Ubi
Posts: 1547
Joined: 17 Jul 2007, 09:01

Re: B3 RAID status problem

Post by Ubi » 20 Feb 2014, 12:04

I apologise for not reading the entire post and responding to all, but i can mention now that the wd black needs to be cooled really well or it dies. Other than that you can replace it.

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem

Post by mcrsro » 21 Feb 2014, 03:27

Ubi,

Thank you for your answer, but I would want also to receive an advice from the B3 designers.

I am particularly interested to obtain from them a clear statement, because I believe they didn't put on sale their product without making a comprehensive house test (putting the product to big "torture" to see the behaviour under different situations).
This machine is very convenient for me, because it is the place where I have deposited all my information I am using daily.
I've read that the IntelliPower created many problems because of the power switching cycles.
Some times it is more efficient to keep continuously a rotating device instead to put it on start/stop cycles, because during the start time it will require a bigger current and, therefore will develop more heat.
At the first glance 4 W more would not increase significantly the heat develped by the HDD inside the B3 or S1 boxes. However I would want to have their opinion.

The problem mentioned by me might be common to other B3 users, because a classic HDD has a short life compared with the life of SSDs. Depending the user expectations, we would want to use B3 for a longer period of time (much longer than 2 years), because it is a very good product. The problem is created by the classic HDD technology used inside.

This could easily overcome if the disks are replaced with SSDs, even that the server CPU is not fast. An SSD would have considerable lower power consumption and, looking for a longer product usage, it would be more convenient to see Br with no failures in 8 to 10 years of continuous functioning.

I just learned that there are 1 TB SSDs available on Amazon at only 500$
They were put on tests and provided a life expectance of more than 10 years under a heavy test cycle, by far havier than a normal regime from B3.

Somebody would tell me that the price for SSD is too big, very near to B3 price itself, but if you want to avoid the RAID system recovery, a solution with SSD RAID would be rock solid for 10 years. Simply configure and forget it!

Once again, thank you for your post, I'm still waiting for a B3 "fathers" answer.

Cristian

johannes
Posts: 1469
Joined: 31 Dec 2006, 07:12
Location: Sweden
Contact:

Re: B3 RAID status problem

Post by johannes » 21 Feb 2014, 14:47

mcrsro wrote:1. It is allowed to replace the 2 TB WD20EARS-00M ("Green Series") with the model 2 TB WD2002FAEX ("Black Series")? I would wish to tell me if the power supplies from B3 and S1 module may support the model WD2002FAEX.
Yes, everything is allowed but this is not recommended, as Ubi points out they will get very hot in the B3 enclosure, which they do not handle very well, causing a shorter probable lifetime. The passive cooling in the B3 is designed for not-so-power-intensive drives, such as WD Green; WD Red or Seagate Green. To answer your question, yes, the power supply will manage it unless you also use the USB ports to warm your coffee.
mcrsro wrote: 2. Western Digital offers other model with 4 TB, having the same power consumption characteristics like the version with 2 TB. Is it possible to use this type? In this case, is it possible to configure the RAID as a 4 TB mirroring assembly?
We have not officially tested and approved any 4 TB drive, even though this should work nicely. So this is for you to try, on your own risk. We have approved several 3 TB drives though.
mcrsro wrote: 3. What is the best succession of operations which can restore the RAID functionality? I mean, is it a better procedure to restore the data from the existing good disk from the RAID ( it is the HDD mounted into S1 module), avoiding to backup the data, but avoiding too the data loosing? Please describe me the most recommended succession.
I would recommend to buy one other 2 TB to replace the bad one with, and do a restore. Speaking in more general terms, consider this:

- never rely on RAID as your only backup of important data. RAID is designed to keep a good up-time (minimizing downtime when disks fail), not to keep data safe. Remember that RAID does not protect you against accidental file deletion, software malfunctions (which could be a virus on a laptop etc.), theft, fire, flood, lightning etc etc. - only disk malfunctions. So please do a proper backup, preferrably off-site, if you have important data.

- If using raid, note the bathtub shaped error rate curve, electronics fail mainly when they are young or old, not in between. Also errors tend to come in pairs, it is more likely that two products next to eachother on the shelf carry the same problem, than two randomly selected products during a product lifetime. Hence, don't buy the two disks at the same time, it would be better to keep the working one you have (at least for another year or so) and replacing the bad one.

Good luck!
/Johannes (Excito co-founder a long time ago, but now I'm just Johannes)

Gordon
Posts: 1349
Joined: 10 Aug 2011, 03:18

Re: B3 RAID status problem

Post by Gordon » 21 Feb 2014, 15:49

I agree. RAID is overestimated as a means to guarantee data safety. It is not. RAID does not eliminate the need to keep backups, and in fact the way that RAID is usually implemented makes it very vulnerable to failure indeed. That is because RAID is designed to be cheap, but the notion that you cannot use the full length of every disk if you use different disks introduces the sense of expense. As a result every RAID is stacked with the same disks that were manufactured about the same time and have a large chance of failing about the same time.

To answer the original question: there have been some posts before where it became obvious that you will need active cooling when using faster disks inside the B3 enclosure. As a plug-in replacement the red disks appear to be the best choice, since the green disks will kill themselves at some point as a result of their constant parking of the heads - which can be fixed but requires manual intervention to make the green disk behave like a red disk.

As far as upscaling is concerned, yes you can replace the failing disk with a larger disk and let the RAID restore itself while not using the full capacity of the new disk. You could then replace the older disk with a larger disk as well and let RAID restore again and then grow the LVM. Whether that's a better solution than starting from scratch and restoring a backup? I doubt it.

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem

Post by mcrsro » 21 Feb 2014, 16:57

Johannes, Gordon and Ubi,

I wish to thank you for your advice.

Gordon, it's a very good idea to go with the Red line of WD. I will try to make a complete backup and to buy two 3GB Red Line HDDs (WD30EFRX), which have 3 year warranty and more storage capacity.
I already found an offer in my country around $170/disk (I am now abroad, but soon I'll be at home to fix the problem of my B3).
I am confident that it's better to use two new devices instead to replace only the defective disk.

I had in the past WD HDDs and, even the warranty was for 3 years, they continue to work after 5 year service! Same time, for everyone, I do not recommend Seagate brand due to many problems I had with all their disks, starting form 2.5" (internal or USB) to 3.5" types.
I will inform you about how it worked!

Kind regards,

Cristian

Ubi
Posts: 1547
Joined: 17 Jul 2007, 09:01

Re: B3 RAID status problem

Post by Ubi » 22 Feb 2014, 06:25

Well said indeed. The thing that may also be worth mentioning is that RAID in the Bx will not improve uptime when the primary disk fails, because the root partition is not mirrored. And when improving uptime is the only real benefit for RAID setups, I keep on wondering why people implement this on the Bx.

johannes
Posts: 1469
Joined: 31 Dec 2006, 07:12
Location: Sweden
Contact:

Re: B3 RAID status problem

Post by johannes » 22 Feb 2014, 07:51

Ubi wrote:Well said indeed. The thing that may also be worth mentioning is that RAID in the Bx will not improve uptime when the primary disk fails, because the root partition is not mirrored. And when improving uptime is the only real benefit for RAID setups, I keep on wondering why people implement this on the Bx.
It may be a bit harsh to say that uptime is the only benefit, but yes, we had some internal debate on this, and we (developers) all agree on the overstated use for RAID, especially in a NAS setup. However, the entire world was against us and customers screamed for RAID, this was back in the B2 days, so we felt we should do it anyways. These argiments are easier won on this forum, where people understand this stuff, but when discussing with purchasing managers on retail chains, it's a different story.
/Johannes (Excito co-founder a long time ago, but now I'm just Johannes)

Ubi
Posts: 1547
Joined: 17 Jul 2007, 09:01

Re: B3 RAID status problem

Post by Ubi » 22 Feb 2014, 13:42

Fair enough. This does actually brings to mind how the hitchhikers travel guide describes the Marketing Department of the Sirius Cybernetics Corporation...

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem

Post by mcrsro » 22 Feb 2014, 14:35

To all who helped me:

The stabborn facts show that my initial decision to buy a B3-S1 assembly (and not a simple B3) was the correct one.
During 2 years and 7 months, the system worked properly keeping my data safe and readily available.
The system worked almost continuously, with small periods when I was not at home and I stopped it.

Now I have a system that still have my data available, and thanks to B3 designers, I will be able to save my data and restore the system functionality, after disks replacement, as it was before.

I am guided by the net result: data are protected better if you have a simple RAID 1 system, as long as the system health is monitored.

Kind regards,

Cristian

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem -help required

Post by mcrsro » 04 Mar 2014, 09:07

Johannes,

1) I bought two disks (3TB Red Line WD30EFRX), and, after I saved all my data I passed to the change of the faulty disk.
2) Followed the indications from the B3 manual to replace the faulty disk.
3) Formatted a flash USB stick as bootable FAT32 and copied the installation information as indicated on the material "Re-installing your Bubba Server / Bubba2 / B3 - ExcitoWiki"
The net result is that after some time when the lED was green, it turned to blinking purple and remained in this status (I don't understand this, because it is not explained in the manual).
4) I thought that probably the boot loader had to be updated (as I have seen in the material:
"Updating your B3 to support 3TB disks", from ExcitoWiki) and I followed the indicated procedure:
> creating another USB flash stick with "uboot_3TB" update
> booting the mashine (LED was green) and keeping on for 2 minutes. After that time I removed the power
> put the Install USB stick, keeping pressed the reset button and applying the power to start B3 reinstallation
> removed the power
5) Hoping that the 3TB disk will be accepted, started the re-installation. I put the installation stick in one of the USB ports, pushed and kept pressed the Reset button and powered on B3. The LED became green, stayed some time in this status and further became blinking purple

The net result after waiting 45 minutes is that the LED still stays in blinking purple.

I have no possibility to know what is inside B3, because it is not accessible.

Could you, please, guide me how to overcome the actual situation? I didn't change the S1 disk with the new one yet; it still contains the RAID member 2.

Thank you in advance

Cristian
Last edited by mcrsro on 05 Mar 2014, 05:22, edited 4 times in total.

Ubi
Posts: 1547
Joined: 17 Jul 2007, 09:01

Re: B3 RAID status problem

Post by Ubi » 04 Mar 2014, 09:30

just a quick question: did you perform the activities decribed in that last post with the S1 removed?

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem

Post by mcrsro » 04 Mar 2014, 09:46

Yes Ubi, because I 've seen that the system must be installed independent of S1.

I already ordered an USB-to-SATA external rack for 3.5" HDD and to check if the new disks from Red Line work properly. This will come tomorrow evening. If the disks ar OK, it might be a problem with the B3 board (don't know what it could be).

What produced me a big frustation is the lack of any tool to access the B3 status, at least to see at what stage reached the B3 software reinstallation.

As a last solution (and only if it would be necessary), how long would take the shipment of a new B3 board by DHL? My B3 is out of warranty period, so I would also want to know (just in case) how much should I pay for this.

Cristian

stasheck
Posts: 107
Joined: 15 Jan 2014, 13:13

Re: B3 RAID status problem

Post by stasheck » 05 Mar 2014, 10:00

Don't worry about your B3, I remember I also failed to install B3 on first try. I'm trying to remember what I did to fix it, I think I just used another USB drive.

mcrsro
Posts: 9
Joined: 19 Feb 2014, 14:56

Re: B3 RAID status problem

Post by mcrsro » 05 Mar 2014, 10:27

Stashek,

I already used 3 USB flash sticks, one being the original USB stick that came with the B3 delivery.

The symptoms were the same: after some time when the LED was green (signalizing an installation process) it turned to blinking purple (which is not explained anywhere in the documentation - last version I downloaded 2 days ago).

My supposition is that, in spite of all B3 "fathers" efforts to implement the 3 TB use with B3, some disks (including WD30EFRX disk) do not work. Therefore, I ordered a 2 TB disk (WD20EFRX) which is also from the Red line (the Green line is no longer available in my country, being considered obsolete) and I shall try again the installation. As you can see, I am trying to fix the problem as much as possible by my own effort.

If this will fail again, I will seriously need the B3 "fathers" help, because otherwise it would be very difficult for me to understand why a desktop computer I have can live 7 years without changing components and a machine, much more simple, like B3, cannot ...
Till now no answer from them and my time (unfortunately I have to leave again my country very soon) is running out!

Anyway, thank you for your help

Cristian

Post Reply