Drive Troubleshooting Flowchart
18 minute read.
Use the flowcharts here to help you diagnose your drive related problem(s). You will have a definitive answer much sooner than if you were to ask the forum for help and waiting on a reply, followed by asking questions for more data, running tests, etc. Save yourself some time and learn a little bit in the process of using these flowcharts.
These are very simple to follow flowcharts to diagnose drive problems/failures. There is a small bit about ZFS problems as many people will say it must be a drive failure prematurely when it is actually a ZFS issue, so a very small part of ZFS is included, this is not a ZFS troubleshooting guide as ZFS troubleshooting can get quite involved.
While the flowcharts themselves are not harmful, misuse of some of the commands provided can cause loss of data. If you do not know what these commands do, then you should not be using them. All are safe if used “exactly” as written. This means that it is up to you to ensure the commands are used correctly. If you endeavor down this path and are unsure about a command, just ask someone on the forums for some assistance.
This is a troubleshooting guide for the most common problems. It is NOT a guide to fix every little problem that can arise from a file system issue, this is focused on physical drive issues. I did include some basic ZFS troubleshooting because many people think these problems are drive failures and this will identify if it is a drive failure or not. Pay attention to the little things that catch your eye, as said, not everything is in these flowcharts but I tried to include the most common seen problems.
You MUST be a privileged user such as root or you may need to use sudo. If you enter a command and get the response that the command cannot be found, then you do not have the privileges to run that command.
Refer to Appendix A for examples of Self-Monitoring, Analysis, and Reporting Technology (SMART) and Field Access Reliability Metrics (FARM) screen outputs and Appendix A also has Amplifying Information for the Chart questions.
Refer to Appendix B when an operation is requested, such as reading SMART data or performing a SCRUB, for what the command is and how to use the command. If you are uncomfortable executing any of the commands, then seek additional help.
Seagate drives may (generally do) report Error Rates differently, it looks like a wild rapidly changing number. Be aware of this for Seagate drives. Often people see this as an error and jump to an incorrect conclusion, especially when experiencing ZFS errors. This guide will help reduce going down the wrong path by using a more systematic approach.
Values I used for Good/Bad evaluations are my personal values that I go by, not some industry standard. Example: “ID5 < 5” means this is my personal value where I allow a few reallocated sectors but when it becomes greater than 5, I replace the drive to circumvent a complete failure. Some people may prefer a value of 1 as a failure, a value of 1 to me means to keep a watch on the value, if it starts to increase then that is a problem.
There are four flowcharts:
1. ZFS ERRORS
2. CRITICAL DRIVE ERRORS - This flowchart is for what the author considers critical errors.
3. NON-CRITICAL DRIVE ERRORS – This flowchart is for what the author considers are non-critical errors, however that does not mean they are to be ignored and pushed off. You still need to take action.
4. SUSPECT FOUL PLAY (ALTERED DRIVE DATA) – The Seagate Drive Issue Saga (Not only a Seagate issue)
This set of flowcharts is not designed to hold your hand and take you step by step to perform the troubleshooting, drive replacements, etc. These flowcharts are very easy to follow, and in the text version, are step by step however they will tell you that your drive is bad. It is up to you to know how to replace it.
ANSWER THE QUESTIONS AS WRITTEN. Do not assume I meant something different or you might go down the wrong path.
Use the TrueNAS Guide to perform things like “How to replace a drive”. It is there and is well written.
To check if you have an SMR drive, you need to play Detective on the internet and find out if your drive is SMR or CMR. If you have SMR, this could be your problem.
These flowcharts will not specifically help you put your pool back ONLINE, it could happen though. It will help with correcting the common ZFS problems seen, which many people attribute to a drive failure.
If you need help, post the entire output of
smartctl –x /dev/???for interpretation. Do not hold back any data (serial number exception), thinking you know better. No one likes to ask for data more than once.If you have recommended changes, reach out to me. I will evaluate it and update if I agree.
Disclaimer Again: All of the commands in these flowcharts are safe “if” you use them as prescribed. If you find yourself not sure what you should do, STOP! Ask the forum members. No one wants to see a person lose data. That could one of us in the future, it certainly was us in the past. Pain, lots of pain. Use at your own risk.
At this point you can choose to reference the graphical images of the flowcharts, which may look good to some people, or too busy for others to easily manage. Or you can use the step by step guide.
GRAPHICAL - Links
ZFS Flowchart Slide 1 - ZFS Errors
ZFS Flowchart Slide 2 - ZFS Slow Pool
Critical Drive Errors Slide
Non-Critical Drive Errors Slide 1
Non-Critical Drive Errors Slide 2
Foul Play (The Seagate Saga)
Appendix A - How to read a SMART Output and Amplifying Information
Appendix B - Commands to help you
How To Use: Read each block carefully and answer each question. If you need to perform a task such as “Run a Scrub”, refer to Appendix B for the command(s) required to perform the action.
Note: If you get stuck in a loop more than 2 times, exit the loop and seek additional help. Use your brain, repeating the same operation over and over again without any change is a waste of time.
1. Is your pool state "ONLINE" or "DEGRADED"?
Yes - goto step 3
No - goto step 2
2. This flowchart does not assist in getting a ZFS Poll back ONLINE.
It can be very involved and you should seek additional help.
With that said, if you feel you do have a drive failure, proceed to step 3
or Exit this procedure.
3. Is Repaired Data from the zpool status greater than zero (0)?
Yes - goto 14
No - goto 4
4. Do you have Read, Write, or Cksum Errors?
Yes - goto 5
No - goto 9
5. Run a Scrub
6. Did the Read, Write, or Cksum Errors increase? Note any corrupt files?
Yes - goto 7
No - goto 9
7. Have you been through this loop more than twice?
Yes - goto 8
No - goto 4
8. You have a serious problem, backup your data and seek additional help.
Exit this procedure.
9. Run a Zpool Clear on your pool
10. Run a Scrub
11. Do you still have errors?
Yes - goto 7
No - goto 12
12. Restore any corrupt files from a backup.
13. Problem Corrected.
Exit this procedure.
14. Record the names of the corrupt files and delete them.
15. Run a Scrub
16. Do you still have Repaired Data errors?
Yes - goto 14
No - goto
17. Have you been through this loop more than twice?
Yes - goto 18
No - goto 3
18. Are the errors associated ONLY with one drive?
Yes - goto 19
No - goto 20
19. Troubleshoot the drive error using the Critical and Non-Critical Flow Charts.
Exit this procedure.
20. You have a serious problem, backup your data and seek additional help.
Exit this procedure.
1. Is your Used Space Over 80%?
Yes - The pool has limited space to write data and it slows everything down.
Typically, when it hits 90% full, it slows down considerably.
Reduce your pool capacity below 80%.
Exit this procedure.
No - goto 2
2. Do you have SMR drives?
Yes - ZFS does not work well with SMR drives during the write operations.
Replace with CMR type drives.
Exit this procedure.
No - goto 3
3. Question to ask yourself for when you may have noticed the change in speed:
* Did you Reboot?
* Check to see if your NAS is just Busy
* Check SNAPSHOTS
* Check SWAP SPACE
* Seek Additional Help
NOTE: NA = Not Applicable Media Wear Level – Can be reported as ‘Wear Level’ or ‘Percent Used’. Wear level value typically starts at 100 and decreases, Percent Used value typically starts at 0 and increases. Examine the ‘Thresh’, if it is below 50 then a value of 100=new, 0=worn out. If ‘Thresh’ is above 50 then 0=new, 100=worn out. Refer to Appendix A for amplifying information on SMART attributes and values.
1. Can you obtain SMART data from the drive?
Yes - goto 2
No - goto 3
2. Is SMART STATUS – FAIL. Media Error, or Critical Warning?
Yes - goto 3
No - goto 4
3. Back up your data, then consider replacing the drive.
Exit this procedure.
4. Did a SMART Short or Long test fail?
Yes - goto 5
No - goto 7
5. Run a SMART Long Test.
Did it PASS or FAIL?
Pass - goto 6
Fail - goto 3
6. Problem Corrected. Monitor for reoccurrence.
Another failure means the drive is bad.
Exit this procedure.
7. For SSD/NVMe, Is Wear Level value ‘0’ Zero or for HDD is Helium value Low?
Yes - goto 8
No or NA - goto 9
8. You have a serious problem.
You are looking at needing to replace the drive immediately.
Exit this procedure.
9. Is Reallocated Sectors (ID 5) raw < 5?
Yes - goto 11
No - goto 10
10. You likely have a serious problem; this kind of error generally means
media damage or wear has occurred. Seek additional help.
The drive should be replaced.
Exit this procedure.
11. Run a SMART Long Test.
12 Did the SMART Long test PASS or FAIL?
Fail - goto 3
Pass - goto 13
13. Is Spin Retry Count (ID 10) raw > 0?
Yes - goto 14
No - goto
14. Does the drive sound like it is spinning up and down?
Yes - goto 11
No - goto 15
15. This could be a drive or PSU failure.
Check the power before replacing the drive.
Exit this procedure.
16. You do Not have a Critical Problem. Refer to Non-Critical Flowchart.
Exit this procedure.
17. Current Pending (ID 197) raw > 0 or Uncorrectable Sector (ID 198) raw > 0?
Yes - goto 18
No - goto 16
18. These errors alone are generally not considered “Critical”.
You do Not have a Critical Problem. Refer to Non-Critical Flowchart.
Possibly seek additional advice.
Exit this procedure.
1. Can you obtain SMART data from the drive?
Yes - goto 18
No - goto 2
2. You have a serious problem, seek additional help.
Exit this procedure.
3. Is UDMA CRC Errors (ID 199) raw increasing?
Yes - goto
No - goto 4
4. * Are Hardware ECC Recovered or MultiZone Errors increasing and the ONLY errors?
Yes - goto 5
No - goto 6
5. This is not an immediate issue, monitor for increasing counts.
Hardware ECC Recovered (ID 195) raw count does not indicate a typical drive failure.
Exit this procedure.
6. Are there any other errors?
Yes - goto
No - goto 7
7. Hardware ECC Recovered (ID 195) is the drive hardware automatically correcting
the data read, as it should. MultiZone Error (ID 200) sometimes is an indicator
of a problem, if associated errors exist, otherwise usually not an issue.
Exit this procedure.
8. Is Current Pending Sectors (ID 197) raw > 5?
Yes - goto 9
No - goto 12
9. Run SMART Long Test.
Does it pass?
Yes - goto 10
No - goto 11
10. This is okay for the drive right now.
Monitor Current Pending Sectors (ID 197) raw.
If it continues to increase, replace the drive.
Seek additional advice if you desire but provide the full SMARTCTL output for evaluation.
Exit this procedure.
11. You have a serious drive problem.
Replace the drive.
Exit this procedure.
12. Is ID 1 Raw Read Error rate or ID 7 Seek Error Rate your concern?
Yes - goto 14
No - goto 13
13. Seek additional advice.
Ensure you post the complete output of `smartctl –x /dev/xxx` when asking for help.
Exit this procedure.
14. Do you have a Seagate Drive?
Yes - goto 15
No - goto 16
15. Seagate does things differently. If you see wildly changing numbers,
this may be normal for your drive.
Take the Number, subtract 4294967295 (FFFFFFFF Hex), if there was a
whole number remainder then you have a real failure.
- Goto 9
16. This indicates a failure of the armature/head mechanism.
Replace the drive.
Exit this procedure.
With the flood of used Seagate and other brand hard drives in the market in 2024/2025, if you bought a new drive, you should verify that it is actually new. The steps to do that are here and I Highly Recommend you do this check. Unfortunately, only Seagate at this time has this FARM data so this will not work for other drive manufacturers.
Western Digital has WDDA however that may not be more than exactly what is in SMART already and is not covered here as it required additional software that is not installed on TrueNAS.
(Shameless Plug -- Multi-Report V3.15 (and later) has this check built-in.)
1. Check which version of smartmontools is installed. `smartctl`
2. Is smartmontools version 7.4 or greater?
Yes - goto 4
No - goto 3
3. You will need smartmontools 7.4 or later to read the FARM data.
Exit this procedure.
4. OBTAIN SMART DRIVE DATA `smartctl –a /dev/driveid`
5. OBTAIN FARM DATA `smartctl –l /dev/driveid` (-l is a lower case L)
6. From both pieces of data, write down the following data:
a. Serial Number
b. Power On Hours
c. Spindle Power On Hours
d. Head Flight Hours
e. LBAs Read and Written
7. Compare the data between the two reports, is there a large discrepancy?
Yes - goto 9
No - goto 8
8. You are one of the lucky ones, your drive appears to be valid.
Exit this procedure.
9. You Likely have an altered drive.
If your drive was refurbished/previously used, the data may be correct.
If in doubt, seek additional help.
Exit this procedure.
The attribute is defined by its ID number. The human-readable name (e.g. “Reallocated Sectors Count” for ID 5) may vary slightly depending on the software you use.
SATA HDDs, SAS HDDs, SATA SSDs, and NVMe SSDs use significantly different sets of attributes.
Some attributes are manufacturer-specific.
Whatever the format is of the SMART data, it will contain similar values, pay attention to what you are reading.
If you have a question about an attribute, Google for “S.M.A.R.T.” and the attribute name.
SMART Data is not terribly difficult to read and understand. There is SMART data illustrating a typical output for a SATA Hard Drive in the Graphical Charts.
If you have a question, ask for help deciphering the data.
This is amplifying information to the troubleshooting flowcharts. Read the entire flowchart section before jumping into it. You can wait 5 minutes to read this before you begin troubleshooting.
When reading SMART data, take in all of the data for the line you are reading. This is the ID, Attribute, VALUE, WORST, THRESH, RAW_VALUE. When reading SMART values, pay attention to what the flowchart is asking for.
- The ID/Attribute defines the meaning of what you are examining.
- VALUE is a Normalized value on the performance of this attribute.
- WORST is the worst VALUE observed over the life of the drive.
- THRESH is the lowest acceptable limit for VALUE and WORST. If VALUE or WORST for any attribute are below THRESH, the drive is considered failed. If THRESH is 0, the attribute is not considered life-critical (e.g. Power On Hours Count).
- RAW_VALUE (raw) is the non-normalized value. This value is used often for many of the attributes and may not be reflected immediately in the Normalized values. The content of the RAW_VALUE is (in general) vendor-specific.
When evaluating SMART data:
- Example: ID195 - Hardware ECC Recovered – VALUE=200, WORST=200, THRESH=0, RAW_VALUE=40945360 (data from my drive)
When the VALUE or WORST approaches the THRESH value, this is typically a failing indication. In this case, the drive is perfectly fine. However, if VALUE=45 and WORST=(45 or less), AND you have other errors, then this drive should be replaced at an opportune time. If something looks wrong, ask for help if you need to. NEVER WAIT UNTIL THE LAST MINUTE!
- Example: ID1/Raw Read Error Rate – VALUE=076, WORST=064, THRESH=006, RAW_VALUE=40945360 (failure number)?
Error rates change, they go down and they go up. In this example you can see that the error rate dropped to 064 but is now 076. Error rates are calculated over a large number of operations. Seagate drives display these large RAW_VALUE number as this value represents more than a total number of failures. So if you have a Seagate made drive, do not worry about the “Error Rate” fields without just asking someone. The flowchart provides the specific math to perform if required.
If you have a question about what you are reading, ask the question. No one wants you to replace a drive unless it is failing, and does not want you to ignore a drive that is failing. Dang, did I say something like this before?
LINK SLIDES 1 to 6 showing screen examples
Below is a list of common commands for both CORE (FreeBSD) and SCALE (Debian Linux) which help diagnose the possible drive issue. Unless specifically identified as CORE or SCALE, the commands work for both. These commands are used in the troubleshooting procedures. Some commands can be dangerous and I will bold those in RED font . But can be simply means, don’t misuse the command, and be very precise and press the correct keys.
FORMAT: Each command will be surrounded by an apostrophe (`), the same way we post commands on the TrueNAS Forum. Blue Font represents the name of your pool, so replace pool with the name of your pool. Orange font indicates your Drive Ident. Green font is Extra Information.
SAFETY OF COMMANDS: These commands are all safe to use as outlined below. Some of the commands have do have destructive power, however you would have to significantly deviate from the examples provided.
zpool status -vProvides the pool status for all pools.zpool scrub poolStarts a SCRUB operation on the selected pool.zpool clear poolClears all Read, Write, and Cksum errors for the designated pool.
Note: Drives may have multiple partitions so when cross referencing by GPTID, you just need to look for a Drive ID, not the different partitions. Drive ada0p1 is the same physical drive as ada0p2. Example “gptid/d0f8a4fe-bf79-11ed-a0df-000c296fd555 N/A ada0p2” This is Drive ID ada0 partition 2.
lsblk -o +PARTUUID,NAME,LABEL,SERIAL[SCALE]: Provides a listing of your disks, partitions, drive ident, and drive serial numbers so you can cross reference all these identifying areas so you replace the correct drive using the serial number.glabel status[CORE]: Provides the GPTID and Drive Ident. “nvd0” = “nvme0” To obtain the Drive Serial Number, use the “OBTAIN DRIVE SMART DATA” section to cross reference the DRIVE IDENT to the DRIVE SERIAL NUMBER.
smartctl --scan[CORE/SCALE]: Lists the interface types for all available drive. If running the smartctl commands below, there is an error reading the drive, try adding the interface type.Format:
smartctl -d interface_type command stringExample:smartctl -d scsi -a /dev/sda
smartctl -a /dev/sda[SCALE]: Provides a BASIC listing of the drive SMART data. Most diagnosis can be made using this data. The Orange indicates the Drive Ident.smartctl -a /dev/da0orada0[CORE]: Same as above.
smartctl -x /dev/sda[SCALE]: Provides EXTENDED listing of the drive SMART data.smartctl -x /dev/da0orada0[CORE]: Same as above.
smartclt -l farm /dev/sda[SCALE]: Provides Field Access Reliability Metrics log ‘FARM’. This data can be useful in determining if a drive has had the SMART data reset.smartclt -l farm /dev/da0orada0[CORE]: Same as above.
SHORT TEST
smartctl -t short /dev/sda[SCALE]: Run a “short” SMART test on the drive.smartctl -t short /dev/da0orada0[CORE]: Same as above.
LONG TEST
smartctl -t long /dev/sda[SCALE]: Run a “long” SMART test on the drive.smartctl -t long /dev/da0orada0[CORE]: Same as above.
STOP TEST
smartctl -X /dev/sda[SCALE]: This will abort the current SMART test if one is running.smartctl -X /dev/da0orada0[CORE]: Same as above.
These commands if misused could be destructive, enter as written, NO typos.
In the examples below, the drive is nvme0, replace this value for other nvme drives.
SHORT TEST
nvme device-self-test /dev/nvme0 -s 1[SCALE]: Run a “short” SMART test on the drive.nvmecontrol selftest -c 1 nvme0[CORE]: Same as above.
LONG TEST
nvme device-self-test /dev/nvme0 -s 2[SCALE]: Run a “long” SMART test on the drive.nvmecontrol selftest -c 2 nvme0[CORE]: Same as above.
STOP SMART TEST
nvme device-self-test /dev/nvme0 -s 0xf[SCALE]: This will abort the current SMART test if one is running.nvmecontrol selftest -c 0xf nvme0[CORE]: Same as above.