S352BU33RER RAID1 operational details

Cards, Drive Accessories, Tools
1

StarTech Community:

I am trying to understand the RAID1 operational details of the USB 3.0 / eSATA Dual-Bay Trayless 3.5” SATA III Hard Drive Enclosure with UASP - 2-Bay SATA 6 Gbps Hot-Swap HDD Enclosure, Product ID S352BU33RER.

Assume all HDD’s have the same number of sectors. Assume that the enclosure is powered down whenever HDD’s are installed or removed:

  1. Given one HDD with content and one zeroed HDD, how can I create a RAID1?

  2. What is the enclosure’s policy for marking a HDD as failed and kicking it out of the RAID?

  3. Given two HDD’s operating in RAID1 mode and one HDD fails, what happens? Will the OS be able to read and write? How can a remote sysadmin know that one HDD has failed? How can a local sysadmin identify the failed HDD?

  4. Given two HDD’s operating in RAID1 mode and one drive is removed, what happens? Will the OS be able to read and write? How can a remote sysadmin know that one HDD is missing? How can a local sysadmin identify the mssing HDD?

  5. Given RAID1 mode with one good HDD and one failed or removed HDD, and a zeroed HDD is installed in its place, what happens? Will the OS be able to read and write? How can a remote sysadmin know that the RAID is resilvering or not? How can a remote sysadmin measure the progress of resilvering? How can a remote sysadmin know that resilvering completed?

  6. Given RAID1 mode with one good HDD and one failed or removed HDD, and a previously removed but good HDD is installed in its place, what happens? Will the OS be able to read and write? How can a remote sysadmin know that the RAID is resilvering or not? How can a remote sysadmin measure the progress of resilvering? How can a remote sysadmin know that resilvering completed?

TIA,

David

2

Hello @dpchrist,

Thank you for joining the StarTech.com community.

I will be happy to take some time to answer your questions as best I can. The support materials on the StarTech.com website may answer some of these questions.

S352BU33RER Drivers & Downloads - USB 3.0 / eSATA Dual SATA HDD Enclosure - External Drive Enclosures | Hard Drive Accessories | StarTech.com

The manual is listed on the drivers and downloads section of this product page as:
s352bu33rer_manual_rev 1.pdf

I will answer your questions.

  1. The S352BU33RER is a hardware RAID solution that manages the installed hard drives. This means that the enclosure will format and partition the drives during the RAID steps. Any valuable data on the drives will need to be backed up prior to installing the drives in the enclosure. We recommend closely following the RAID setup steps in the manual that is linked above to create a new RAID1 volume. Once the RAID volume is created, you can then transfer your backed up data on to the new RAID volume.

  2. If a drive in the S352BU33RER has failed the rebuild LED will illuminate along with the drive LED. This is an internally managed device, and the specific policy for determining a failure is managed by the onboard hardware. I anticipate that this will include standard SATA or SMART errors.

  3. If the S352BU33RER is operating in RAID1 mode and one drive fails, you will continue to be able to use the enclosure normally. I suspect there may be some deterioration in performance. We would strongly recommend testing and replacing the failed drive as soon as possible. Since the S352BU33RER is a hardware based solution that is recommended for use on a local system, no management software is required so there is no way to monitor the health of the RAID remotely.

  4. If one drive is removed from the S352BU33RER the enclosure will continue to function but the rebuild light will be illuminated indicating an error. There are no remote administration features available for this enclosure.

  5. If one drive is wiped and replaced, the S352BU33RER will initiate a rebuild. The rebuild may take some time complete, this will depend on the size of the drives. The rebuild may take some time to complete. The enclosure will be accessible to the operating system but changes may prolong rebuild times.

  6. We do not recommend removing and replacing drives in the S352BU33RER data changes between RAID pairs may result in long rebuild times, or data loss if there are two drives that were previously a RAID pair but one was removed. If drives do need to be removed and replaced for any reason, it is best to label the drives to be sure that they are placed back into the original bay.

Another note to keep in mind, as mentioned in the manual, that if you are switching between modes, you may need to configure the S352BU33RER to JBOD mode as a step between changing between different RAID modes. The JBOD mode configuration resets the enclosure.

I hope that this answers your questions, and gives you a better understanding of the capabilities and limits of this enclosure.

Please let us know if you have any other questions, or would like any further assistance.

Sincerely,

@DavidF

3

StarTech:

Thank you for the information. I understand that these are tough questions, but the answers will be critical for creating a disaster preparedness and recovery plan that involves the S352BU33RER.

Please clarify your responses to the previous questions:

2a. What is the enclosure’s policy for marking a HDD as failed and kicking it out of the RAID? E.g. What is the technical criteria – N failed reads in time T, retried M times, etc.? Can this be tuned?

2b. So – if drive 1 fails, the Drive 1 Activity LED is steady on and the REBUILD LED is steady on? And – if drive 2 fails, the Drive 2 Activity LED is steady on and the REBUILD LED is steady on?

3a. When a drive fails, will the OS know this? E.g. Event Viewer, systemd, etc.?

3b. Does the S352BU33RER support monitoring via SMART or any other API? E.g. smartctl(8)?

6a. One backup strategy is to create a 2-way mirror of 3 disks, always keeping one disk out-of-mirror in rotation. Start by building a mirror of two blank disks (A and B). After one period of time (week, 2 weeks, month, etc.), remove disk A and insert a third blank disk (C). After resilvering, the mirror contains disks B and C; disk A is out-of-mirror and can be stored off-site. After a second period of time, remove disk B and insert disk A. After resilvering, the mirror contains disks A and C; disk B is out. After a third period of time, remove disk C and insert disk B. After resilvering, the mirror contains disks A and B; disk C is out. Repeat the process indefinitely. If one disk fails at any time, replace it with a blank disk (labeled the same as the disk being replaced). If both disks fail, remove them, insert the out-of-mirror disk, insert a blank disk (next label), and allocate another blank disk for the next rotation (second next label). Will this strategy work on the S352BU33RER?

Some new questions:

  1. If one RAID1 disk is removed and the S352BU33RER is set to JBOD mode, what happens? Will the OS be able to read and write?

  2. Following #7, if the S352BU33RER is set to RAID1 mode and the previous second RAID1 disk is inserted, what happens? Will the OS be able to read and write?

David

4

Hello @dpchrist,

Thank you for getting back to us with your clarifying questions. Your questions are not unreasonable, however I do want to be sure that you understand the capabilities of the S352BU33RER so that you can determine if it will meet your requirements.

I may need to inquire further with my team to specific details. However, based on your line of questioning, I am beginning to suspect that the S352BU33RER may not be the right product for you.

The S352BU33RER does provided feedback with LEDs, and there is some level of SMART data available when it is used in JBOD mode. However, I do not see any features that indicate whether it will notify a failure of a RAID array using SMART data. If you require features that will allow you to administer the device remotely and perform more advanced backup procedures, like the ability to reliably swap drives in and out, I can certainly see if we have another solution that may be a better fit.

I will be happy to reach out to my team for more, but a few more details may help us understand how to direct this request.

  1. Are you currently working with the S352BU33RER or are you interested in purchasing?
  2. How many S352BU33RER do you anticipate purchasing, or how many units do you have?
  3. Is this request for a company? Are you comfortable sharing the company name? If you wish I can reach out to you more privately with the email that you used to sign up for the forum.
  4. Can you provide more detail about why you are interested in the S352BU33RER, specifically?

Please add any details you can, but if you prefer I can take a moment to open a case for you and interact directly.

Once I have these details, I will be able to investigate this more deeply to see if this information is available to us.

Sincerely,

@DavidF

5

StarTech:

Thank you for responding. Here are the answers to your questions:

  1. I have purchased one S352BU33RER for deployment at a client site with a small WFW/SMB network of Windows 10 computers. We want to implement backups using File History and Backup and Restore (Windows 7) by connecting the S352BU33RER to one computer, sharing one or more directories over the network, and configuring all of the computers to backup to the network share(s). On-site/ off-site rotation of backup drives is required.

  2. I have one S352BU33RER now. I may buy/ recommend more in the future, depending upon how this project works out.

  3. I am a freelance consultant. Keeping this thread on this forum benefits everyone.

  4. I selected the S352BU33RER because it supports two 3.5" SATA III drives, has trayless racks, has a cooling fan, has both eSATA and USB 3 interfaces, and supports various operating modes; notably dual independent and RAID1. (The client does not want a NAS.)

David

6

Hello @dpchrist,

Thank you for answering my questions.

You are definitely asking some very specific questions, and we might not have the granularity you are looking for, but I will do my best to gather that for you.

A RAID1 enclosure like the S352BU33RER is very good high availability data solution, and it can certainly be part of a backup solution. However, based on my understanding, I am not optimistic that it will be a reliable solution for the type of backup scenario outlined here.

To be sure that I am understanding this product correctly, and providing the correct answers, I have asked my team for their input. I wish to provide you the best information to confirm if the S352BU33RER is the right solution for you.

I will reply here again once I have heard back from them.

Sincerely,

@DavidF

7

Hello @dpchrist,

I have heard back from my team. As I had suspected, many of these questions are outside of the scope of this product, but I will try my best to point you in the right direction if this is information that is not available to me.

My team highlighted for me is that S352BU33RER is designed to be a simple 2-bay RAID enclosure. If more advanced features are required you may need to seek another solution. I did take a moment to review our currently available hard drive enclosure but I do not see another option with more advanced features I can suggest.

The best way I can follow on with your questions from #2, is to confirm that the S352BU33RER uses a Jmicron chipset. The parameters like RAID rebuild, error criteria, notifications, etc. will all be integrated and managed by this chipset.

The chipset information for many of our products can be found in the Technical Specifications section on our product pages. This may help you find some more clues:

S352BU33RER Technical Specifications - https://www.startech.com/en-us/hdd/s352bu33rer#technical-specifications

Further to #3. My colleagues did not have specific information about RAID errors or SMART data but noted that some limited logging could appear in the operating system logs. Event Viewer in Windows, the Console logs in macOS, or the system or kernel buffer in Linux. SMART data would be limited particularly in RAID modes. I apologize that we do not have more specific information to offer.

My team confirmed that the LEDs on S352BU33RER are the expected and recommended way to monitor the function and state.

Regarding #6, the 2-way mirror backup strategy of keeping one disk out-of-mirror in rotation. We cannot recommend this approach. We only recommend rebuilding the array only when truly necessary. My team encourages backing up the data on the RAID volume to another drive or storage solution at regular intervals.

#7 - My colleagues confirm that we have seen a single drive RAID1 continue to operate when set to JBDO mode, however, the data in the RAID1 mode may likely be lost if it is switched back to RAID1. We only recommend changing the RAID mode if the data is first backed up and when the previous RAID configuration is no longer needed.

#8 - Placing a previously installed RAID1 drive where another RAID1 drive is already in place may lead to data loss. We do not recommend swapping RAID drives.

Again, we encourage you to consider a solution like the S352BU33RER in RAID1 mode as a high-availability data solution that can continue functioning if a single drive fails. It is not designed to be a self-contained backup solution. A desktop RAID enclosure like the S352BU33RER is intended to provide redundancy and minimize downtime from disk failures, not to act as a backup or prevent data loss entirely. Always back up any important data on your RAID separately to ensure no data loss occurs. The best backup solutions will be periodic incremental backups of your important data in multiple locations.

I appreciate that many of these answers do not directly answer your questions and I apologize for the frustration. You are looking for a solution that will meet your particular scope but I am describing the data we have for the S352BU33RER at the scope of the product. If I can be candid, it sounds like the S352BU33RER is not a fit for you. I prefer to offer advice that ensures a best fit and a good experience. The S352BU33RER would be a great fit for someone who wants reliable storage or higher speed and capacity on their desktop lab or studio. It would be apt for photos, video, or other complex media and applications with greater needs than is available with standard external media devices.

Please let us know if you have any other questions, and we will do our best to help you.

Sincerely,

@DavidF

8

StarTech:

Thank you for the additional information.

Your WWW forum software is messing with item numbering. I have enclosed previous item numbers in paratheses to protect them.

I did some research and testing over the past few days using Debian 11.9 Linux 5.10.0-28-amd64 and three (3) Seagate ST31500341AS 1.5 TB 3.5" SATA II hard disk drives (identical sizes, firmware versions CC1H and CC3H):

(1.) The surest way to create a RAID1 is to insert two blank disks. During testing, the S352BU33RER hid the last 97,072 sectors on both disks and wrote metadata there. This process completed within 1 minute. The OS saw one blank disk “SSI H/W RAID1” whose size was smaller than the size of the raw disks.

(4.) When one disk was removed from an operating RAID1, the Drive Activity LED for the removed drive turned off. Contents of the RAID1 were preserved and the OS could read and write. This was tested both hot-plug and cold-plug for both drive bays.

(5.) When the RAID1 was operating on one disk and a blank disk was inserted, a rebuild was initiated. Both Drive Activity LED’s and the REBUILD LED were flashing during rebuild. Contents of the RAID1 were preserved and the OS could read and write during rebuild. Rebuild time was approximately equal to the time required to perform a sequential copy of all sectors from one disk to the other (~4.5 hours). When the rebuild completed, the REBUILD LED turned off. This was verified cold-plug for both drive bays.

(6.) When the RAID1 was operating on one disk, a previously removed disk was inserted, and there were no changes to the RAID1 content while the removed disk was out, the RAID1 resumed operation using both disks. If the content of RAID1 changed while the removed drive was out, a rebuild was initiated (as above) and the inserted disk was overwritten. This was tested cold-plug for both drive bays.

(3b.) smartctl(8) reports for the RAID1 provided a modified subset of the information provided for one disk (top bay?). smartctl(8) short tests generated an error message “scsi error badly formed scsi parameters”. smartctl(8) long tests hung the terminal. Ctrl+C did not interrupt the process – kill(1) was required.

(6a.) A backup strategy using the S352BU33RER in RAID1 mode with a three or more disks in rotation (two live, the rest out) appears to work as desired.

David

9

Hello @dpchrist,

From these results it sounds like the S352BU33RER otherwise operating as expected. However, please be be extra cautious with this backup strategy, consider a separate backup solution. I am concerned that there may be data loss if there is a disruption.

Regarding the numbering on the forum, the numbered list function allows you to start a new line with a number then every subsequent new line starting with a number after that will automatically increment, it does not matter which number is used on subsequent new lines. If5,

I sure do appreciate you posting your testing results here. We hope that the S352BU33RER continues to operate correctly for you.

Please let us know if there is anything else that we can do to help in the future.

Sincerely,

@DavidF