Troubleshooting
A collection of things any LTO user once in a while runs into. As with all things hardware, from time to time, there will be errors.
"Tape changed to read-only by LTFS"
This error is LTFS's way of reporting something is wrong with either the tape or hardware.
Whenever LTFS drops a tape to Read-Only mode, it does so because too many write errors are reported by the writing part of the tape head in the drive. This results in Canister having to fail the Archive, as there's an external issue that needs solving.
Possible causes
Faulty tape media (error code
6)Faulty/dirty/worn LTO drive head (error code
5or6)Out-of-date LTO firmware
A too-high operating or environmental temperature
When it comes to troubleshooting, the steps below should be followed in order:
Check the Single Character Display (SCD) on the front of the drive. Sometimes all you need to do is insert the cleaning tape, but only if you see a
C. Other codes are documented below.If the LTO drive is running on old firmware, perform the relevant upgrade and try again. The success rate of this treatment is very high, so don't skip it.
Try with another tape, ideally from a different batch. Every tape surface comes with irreparable faults, with faults often recurring within a batch, so make sure to rule this out.
If the problem persists, the next step is to request a diagnostic from the drive vendor.
My LTO device displays an error code. What does it mean?
Every LTO drive is essentially a computer, and as such follows its own instructions and error handling. Most drives have a single-character display (SCD) to tell you how it's doing. It's behind a black see-through panel so you won't see it except when booting (a countdown shows) or when an error is stated.
Anytime there is a number or letter visible on the front of your LTO drive, the software is not able to communicate with the drive. Refer to this list of SCD error codes to find out how to resolve the issue at hand. As these are hardware errors, get in touch with your vendor to sort out the issue if needed.
SCD Code | Description |
0 | No Error. Ran successfully. |
1 | Cooling problem. |
2 | 5V dc power problem. Tape drive detected that the Drive Power Supply is approaching the specified voltage limit (drive is still operating) or is outside the specified voltage range (drive is not operating). |
3 | Tape drive determined that a microcode error occurred. |
4 | Microcode or tape drive problem. Tape drive determined that a microcode or tape drive hardware failure occurred. |
5 | Tape drive problem. Tape drive determined that a hardware failure occurred. |
6 | Tape drive or media error. Tape drive determined that an error occurred, but it cannot isolate the error due to faulty hardware or to the tape cartridge. |
7 | Media error. |
8 | Tape drive, SCSI bus or fibre channel error. |
9 | Tape drive or RS-422 error. |
A | Tape drive hardware problem. |
B | No error or message is assigned. |
C | Tape drive needs to be cleaned. |
D | No error or message is assigned. |
J | A too-new tape is loaded into an older-generation LTO drive. |
= | The Unload button on the drive was pushed in and did not release. |
"Medium consistency check failed"
This error means LTFS has detected the last known tape index is faulty, and thus cannot mount. If mounting does not succeed due to index issues, LTFS will itself do a recovery, and then throw this error if that's unsuccessful.
This error is often caused by a disconnected Thunderbolt cable or a power loss. If this tape has no valuable data on it, it's by far the easiest to just reformat it. If that's not possible, you can try a Repair action in Canister.
Canister's Repair option triggers a full recovery, which can also be done directly in Terminal:
After a full recovery, a _ltfs_lostandfound folder is created on the tape. LTFS has created binaries of all non-referenced data and added them to the index. This is not really useful stuff, but it's an indicator that a rollback is needed to get to the data.
It's also possible to do a deep recovery. First unmount the tape, then do…
If none of this helps, it's time for a rollback.
Rollbacks
For each change on a tape, LTFS creates a new index. These are called "generations." Consider them snapshots, and it's possible to roll back to a previous snapshot. When you find a snapshot that works, you can continue with that and archive new data to tape.
In Terminal, list the rollback points. Note: this process takes time! Go grab some coffee.
Creating a rollback list is done top down, starting with the most recent one. Monitor the timestamps to find the desired rollback point, and cancel with Command-C.
Then, roll back to that generation:
Here, 57 is the generation number, and 0 is the drive ID. If you have one LTO drive, it's always 0.
When the process is done, mount the tape and you'll see its contents restored.
My tape mounts, but it suddenly contains a lost_and_found folder
If your tape is showing a lost_and_found folder, the tape has been Repaired, only somewhat unsuccessfully.
When a tape cannot be mounted, LTFS will report if a Repair is needed. If the Repair process locates data on the tape's data partition that is not referenced in the index partition, it's added as "lost and found" data. This is raw data and does not expose file names, so for most purposes, it's quite useless. Only if you know what should be on that tape and what is missing you would be able to cross-reference file sizes and rename the files in that folder through Terminal's mv command.
Last updated