Docs · Reference
Error messages
Every message the agent, the install script and the platform can print, with its cause and the fix. The list is generated from the source code. Messages appear exactly as the agent prints them in English (`--lang en`, the default); the install script prints English. Platform answers are printed in French with `--lang fr`. `…` stands for a value (a path, a date, a detail).
Markdown versionThe CLI prints English by default (`--lang fr|es` for French or Spanish). Some output blocks come from a French session: they are labelled and translated in the text.
Start here
neverwiped status
tail -n 50 /var/log/neverwiped.logA message after Erreur : (“Error:”) on standard error means the command failed (exit code 1, or 2 for wrong usage). A line starting with Avertissement : (“Warning:”) didn’t stop anything.
All messages
Installing
| Message, as printed | Cause | Fix |
|---|---|---|
… …: no Docker volume of that name in the plan (see the list above) | --pause-volume or --live-volume names a volume that is not in the plan (typo, or a volume of a database already exported). | Copy the name from the plan shown just above, or from docker volume ls. |
installation stopped, nothing was registered. Adjust the plan with --include / --exclude / --no-databases / --no-docker and run “neverwiped install” again | You declined the proposed plan. Nothing was registered. | Run neverwiped install again with --include, --exclude, --no-databases or --no-docker. |
missing API key: add --api-key KEY (create it in your client area, “API keys” section) | --yes without an API key. | Add --api-key nw_acc_… or set NEVERWIPED_API_KEY. |
server registration: … | The platform refused to register the server; its reason follows the colon. | See the platform answers below. |
storage access: … | The storage can’t be reached (outbound port 443 closed?). | Check outbound HTTPS from the server to the storage address. |
creating the encrypted repository: … | restic could not create the encrypted repository. | Run neverwiped install --yes again; if it persists, send neverwiped status to support. |
adding the recovery key: … | restic could not add the recovery key to the new repository. | Run neverwiped install --yes again; if it persists, send neverwiped status to support. |
Warning: plan not sent to the platform: … | The plan was saved on the server but not sent to the dashboard. | Nothing to do: backups run; the dashboard shows the plan after the next neverwiped install --yes. |
reading the subscription: … | The platform couldn’t be asked for the subscription after registration. | Check outbound HTTPS, then neverwiped install --yes again. |
Warning: automatic scheduling impossible: … | Neither systemd nor cron is available. | Install cron (apt install cron) and run neverwiped install --yes again, or schedule the command shown yourself. |
backups are suspended for this account (inactive subscription) | The subscription was inactive at the first backup. | Check your plan in the dashboard. |
Note: a backup started by the systemd timer is stopped after 12 hours. For the first backup, remove the limit (neverwiped limits --limit-upload off) and set it again once it is done. | At the upload limit, sending the plan may take longer than the 12 hours the systemd timer gives a backup. | Run neverwiped limits --limit-upload off, let the first backup finish (neverwiped backup now), then set the limit again. |
platform unreachable (…): … | Network, DNS or outbound firewall. | curl -I https://neverwiped.com from the server. |
unknown preset …: choose among …, or all | --preset names no known preset. | Use logs, build, trash, disk-images, archives or all (see the getting-started guide). |
wrong SHA-256 checksum of restic (expected …, got …): download refused | restic download altered, or a wrong mirror. | Check NEVERWIPED_RESTIC_BASE_URL, run again. |
decompressing restic: … | The downloaded restic archive is damaged. | Run again; with a mirror, check NEVERWIPED_RESTIC_BASE_URL. |
unsupported architecture … | The CPU is neither amd64 nor arm64. | Not supported. |
downloading restic: … | GitHub can’t be reached from the server. | Open outbound access to github.com and objects.githubusercontent.com, or use a mirror with NEVERWIPED_RESTIC_BASE_URL. |
downloading restic: HTTP … (…) | GitHub or the mirror answered with an error. | Same as above. |
unsupported schedule: … | The platform sent a schedule this agent doesn’t know (older agent?). | Reinstall the latest agent with the install line from the dashboard. |
neither systemd nor cron is available: schedule “… backup --scheduled” yourself (…) | Neither systemd nor cron is available. | Install cron, or schedule neverwiped backup --scheduled yourself. |
Warning: no cron daemon is running, so this schedule will not run until one is started (…) | The schedule was written to cron, but no cron daemon is running (Alpine does not start crond by default; minimal images often have none). | Run the command in brackets (Alpine: rc-update add crond default && rc-service crond start), then neverwiped status. |
neverwiped install.sh: only Linux is supported. | Not a Linux system. | Use a Linux server. |
neverwiped install.sh: unsupported architecture … (amd64 and arm64 only). | CPU is neither amd64 nor arm64. | Not supported. |
neverwiped install.sh: no published checksum for linux-…; contact support. | The platform has no published binary for this architecture. | Write to support. |
neverwiped install.sh: Linux kernel … is too old (3.2 or newer is required). Nothing was installed. | The Linux kernel is older than 3.2, the minimum of the Go runtime the agent and restic are built with (CentOS 6 and older). | Use a system with Linux 3.2 or newer, for example RHEL/CentOS 7 or later, Debian 8 or later, Ubuntu 14.04 or later. |
neverwiped install.sh: run as root, for example: curl -fsSL …/install.sh | sudo sh | Run without root. | Pipe into sudo sh instead of sh. |
neverwiped install.sh: download failed: … | The server can’t reach the platform. | Check outbound HTTPS (firewall, proxy, DNS). |
neverwiped install.sh: curl or wget is required. | Neither curl nor wget. | apt install curl, dnf install curl or apk add curl. |
neverwiped install.sh: sha256sum or openssl is required to verify the download. | No tool to check the fingerprint. | Install coreutils or openssl. |
neverwiped install.sh: checksum mismatch for neverwiped-linux-… (expected …, got …). Nothing was installed. | The download was altered (proxy, cut connection). | Run it again; if it persists, write to support. Nothing was installed. |
Backups
| Message, as printed | Cause | Fix |
|---|---|---|
usage: neverwiped backup now | backup without now. | Type neverwiped backup now. |
backups suspended for this account (inactive subscription); restore is still possible | Trial ended, payment overdue past the grace period, or cancellation in effect. | Sort it out in the dashboard; backups resume on their own. Restore still works. |
neverwiped is not configured on this server yet: run “neverwiped install” | The agent isn’t installed, or /etc/neverwiped/config.json was deleted. | neverwiped install --api-key … --yes. |
unexpected repository URL … | The repository address in the configuration is malformed. | Don’t edit /etc/neverwiped/config.json by hand; reinstall or run neverwiped recover. |
restic produced no summary | restic stopped early (memory, full disk, signal). | df -h /var/cache /var/lib/neverwiped, free -m, then neverwiped backup now. |
empty retention policy: cleanup cancelled for safety | The platform sent no retention. | Nothing is deleted. Write to support if it repeats. |
Warning: report not sent to the platform: … | The platform couldn’t be reached after the backup. | The backup is done; the dashboard updates at the next one. |
the check found errors in the backups of this server (the NeverWiped team has been told) | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
another neverwiped operation is already running | A scheduled backup or a restore is running. | Wait for it: tail -f /var/log/neverwiped.log. |
the platform refuses the backup: … | The platform refused the backup; its reason follows. | Usually an inactive subscription: see the dashboard. |
nothing to back up: no path of the plan exists | The folders of the plan no longer exist. | Run neverwiped install --yes to recompute the plan. |
Warning: … unreadable file(s), for example … | Unreadable files (permissions, files deleted during the backup). | Harmless once in a while; exclude them otherwise. |
… (storage refused: your plan's quota is reached or the subscription is suspended; see your client area) | Storage full for your plan, or subscription suspended. | neverwiped status (storage line); exclude files or move up a plan. |
Warning: cleanup report not sent: … | Same, after pruning. | Nothing to do. |
unsupported database type: …/… | The saved plan names a database type this agent doesn’t handle. | neverwiped install --yes to recompute the plan. |
Warning: … MySQL/MariaDB table(s) not in InnoDB (…), for example …: they are not in the same snapshot as the rest of the export. Convert them with ALTER TABLE … ENGINE=InnoDB (docs: getting-started#database). | Some tables use MyISAM or Aria. The export takes every InnoDB table from one snapshot without locking your site; these tables are copied as they are when their turn comes. The backup still succeeds. | Convert them: ALTER TABLE db.table ENGINE=InnoDB; (the docs, “A database”, show how to list them all). |
docker not found | Docker was removed since the install. | neverwiped install --yes (or --no-docker). |
empty export | The dump tool produced nothing (empty database or missing rights). | Check by hand: runuser -u postgres -- pg_dumpall > /dev/null or mysqldump --all-databases > /dev/null. |
export cut off: the file does not end with "-- Dump completed" | The MySQL/MariaDB export stopped before its end (server stopped or restarted, disk full, connection lost); the backup is marked failed so a partial dump is never taken for a good one. | Check df -h /var/lib/neverwiped and the database log, then neverwiped backup now. |
volume not found: … | A Docker volume of the plan was deleted. | neverwiped install --yes. |
Warning: Docker volume(s) … backed up live, without the pause copy: … | The copy of a volume with its containers paused could not be made; the volume was backed up live this time, so the backup is there but may hold half-written files. The reason follows. | Fix the reason (next messages), then neverwiped backup now. |
not enough free space for the pause copy (… to copy, … free): free some space, or back this volume up live with “neverwiped install --live-volume NAME” | The pause copy needs the volume's size free on the disk of /var/lib/neverwiped, plus 10 % and 1 GB. | Free space (df -h /var/lib/neverwiped), or, for a big volume written whole, neverwiped install --yes --live-volume NAME. |
the copy took longer than …: copy stopped, containers resumed | The local copy took longer than the pause limit (5 minutes); the containers were resumed so the application does not stay frozen. | For a big volume, neverwiped install --yes --live-volume NAME, or use the application's own export and back that up. |
containers paused for the copy could not be resumed (…): run “docker unpause” on the IDs listed in /var/lib/neverwiped/paused | Docker refused to resume containers the agent paused for the copy. They are still paused: the application does not answer. | Right away: docker unpause $(cat /var/lib/neverwiped/paused), then docker ps. The next backup also tries again. |
Restore and uninstall
| Message, as printed | Cause | Fix |
|---|---|---|
no backup available | No successful backup yet. | neverwiped backup now. |
ambiguous ID … | Id prefix too short. | Give more characters (see neverwiped list). |
no backup matches …; available dates: … | No backup for that date or id. | Pick one of the dates listed. |
the approval code does not match this restore (the backup, the folder or the paths differ from the approved plan); nothing was written. Run the command again with --plan instead of --approve and ask for a new OK | The command given with --approve does not run the plan that was approved: another backup (for example latest after a new backup), another folder or other paths. Nothing was written. | Run the same command with --plan instead of --approve, read the new plan and give a new OK; then run the command it prints, unchanged. |
restore cancelled, nothing was changed | You declined the confirmation. | Run again and type RESTORE, or add --target. |
Note: restore not reported to the platform (…); this does not affect the restore. | The restore worked; only the note for the dashboard failed. | Nothing to do. |
usage: neverwiped recover [--api-key KEY] [--server ID|HOSTNAME] [--recovery-key-file FILE] [--repo-key] [--yes] [--api-url URL] | recover takes no positional argument. | Name the lost server with --server ID or --server HOSTNAME. |
this server already has a neverwiped configuration (…): use “neverwiped list” and “neverwiped restore” here; recover is for a new server after the loss of the old one | This server already has an agent configuration: it isn’t a new server. | Use neverwiped list and neverwiped restore here. Use recover on the replacement server. |
a key file already exists (…): move it away before running recover | A key file from an earlier install is still there. | Move /etc/neverwiped/repo.key elsewhere (keep it), then run recover again. |
missing API key: add --api-key KEY (an account key, created in your client area, “API keys” section) | --yes without an account API key. | Add --api-key nw_acc_… or set NEVERWIPED_API_KEY. |
list of the servers of the account: … | The platform refused or didn’t answer the list of servers; the reason follows. | Check the key (an account key nw_acc_…) and outbound HTTPS. |
this recovery key is not the one of …: it is the key of the server … (ID …). Nothing was revoked. Run “neverwiped recover --server” with that server, or use the key written for … | The recovery key you typed belongs to another server of the account (its check code matches that one). | Nothing was revoked. Run neverwiped recover --server with the server named in the message, or find the key written for the server you chose. |
this recovery key does not match the check code of … (…, shown in your client area next to the server). Nothing was revoked. Check for a typo, or whether the key belongs to another server | The recovery key does not match the check code recorded for this server, nor that of any other server of the account. | Nothing was revoked. Look for a typo (the key is 6 groups of 6 characters), compare the check code on your paper with the one on the server page of your dashboard. |
Note: this is an earlier recovery key of … (check code …), replaced since by the key …. It still opens the backups: recovery goes on with it. | The key is an earlier recovery key of this server, replaced since with recovery-key new. | Nothing to do: it still opens the backups. Keep the newer key too. |
Note: no check code is recorded for … (installed before this check): the key can only be verified after the credentials are re-issued. | The server was installed by an agent older than the check code, so the key cannot be compared before the credentials are re-issued. | Make sure you hold the key of this server before confirming: a wrong key is only detected after the old credentials are revoked. |
Warning: the agent of … contacted NeverWiped on …, less than 24 hours ago. If that server is still running, it is not lost: recovering here stops its backups. | The server you chose was in contact with NeverWiped less than 24 hours ago. | If it still runs, answer no: recovering here revokes its keys and stops its backups. Choose the lost server, or restore files on the running one with neverwiped restore. |
recovery cancelled, nothing was changed | You declined the confirmation. | Nothing changed; run recover again when ready. |
the backups of this server were deleted (server removed from the account, data purged): nothing to recover | The server was deleted from the account and its backups purged. | Nothing can be recovered. Protect the new server with neverwiped install. |
recovery of the server: … | The platform refused the recovery; the reason follows. | See the platform answers below. |
the recovery key does not open the backups of … (wrong key?). Nothing was kept on this server. The storage credential of the lost server was revoked anyway: check the key and run “neverwiped recover” again | Wrong recovery key, or the key of another server. | Check the key you wrote down at install (the one for this server), then run recover again: the old credential stays revoked, a new one is issued. |
the repository of … is empty or does not exist on the storage (the lost server never completed its installation?) | The lost server never finished its install: no backup exists. | Nothing to recover; protect the new server with neverwiped install. |
access granted, but the backups could not be listed: …. The configuration is kept: try “neverwiped list” again later | Access works but listing the backups failed (storage busy or network). | Run neverwiped list again in a few minutes. |
reading the recovery key: … | The file given to --recovery-key-file can’t be read. | Check the path and that you run as root. |
missing recovery key: add --recovery-key-file FILE or set NEVERWIPED_RECOVERY_KEY | --yes without a recovery key. | Add --recovery-key-file FILE (or NEVERWIPED_RECOVERY_KEY). |
empty recovery key | The recovery key given is empty. | Check the file or the variable. |
no server on this account: nothing to recover (to protect this server, run “neverwiped install”) | This account has no server. | Check you used the key of the right account; else neverwiped install. |
no server … on this account; servers: … | No server of the account has that ID or hostname. | Pick one of the servers listed. |
several servers are named …: give the ID with --server (…) | Several servers share that hostname. | Use the ID shown in brackets. |
this account has several servers: choose one with --server ID or HOSTNAME (…) | --yes and several servers on the account. | Add --server ID or --server HOSTNAME. |
invalid choice … | The answer isn’t one of the numbers shown. | Run again and type a number from the list. |
usage: neverwiped recovery-key check [--recovery-key-file FILE] | neverwiped recovery-key new [--yes] | recovery-key needs check or new; --recovery-key-file goes with check only. | neverwiped recovery-key check or neverwiped recovery-key new. |
this key does not open the backups of …. Check for a typo. If you no longer have the right key, make a new one while this server still runs: neverwiped recovery-key new | The key typed in opens none of this server’s keys: a typo, or the key of another server. | Check the label next to the key (server name and check code). If no copy works, run neverwiped recovery-key new while this server still runs. |
the key could not be tested (storage unreachable?): … | The storage could not be reached, so the key was not tested. | Run neverwiped status, then try again in a few minutes. |
no repository key on this server (…): a new recovery key can only be made on the server that holds it. On a new server, use “neverwiped recover” with a recovery key you still have | /etc/neverwiped/repo.key is missing: a new recovery key can only be made with it. | On the original server, put back /etc/neverwiped if you have a copy. On a new server, neverwiped recover needs a recovery key you still have; without one, the backups can’t be read. |
cancelled, nothing was changed | You answered no. | Nothing was changed. |
the repository key of this server does not open its backups, so no key can be added: … | repo.key on this server no longer opens its repository (file changed, or storage error). | Run neverwiped check; if it fails the same way, write to support with the output (never the key). |
adding the new recovery key: … (nothing was shown; your current keys are unchanged) | restic could not add the key (storage unreachable or quota full). | Check neverwiped status and try again. Your current keys are unchanged. |
the new key was added but could not be checked, so it is not shown: …. Run “neverwiped recovery-key new” again | The key was added but the check that it opens the backups failed, so it was not shown. | Run neverwiped recovery-key new again. The unshown key is known to nobody. |
the new recovery key works, but its check code could not be saved in the configuration: … | The new key works and was shown, but config.json could not be written. | Check the disk space and the rights on /etc/neverwiped; keep the key you wrote down. |
uninstall cancelled | You declined uninstall. | neverwiped uninstall --yes. |
a backup, restore or check is running: wait for it to end (neverwiped status), then run uninstall again. Nothing was removed. | A backup, restore or check holds the run lock; uninstalling now would delete its files under it. | Wait for it to end (neverwiped status), then neverwiped uninstall again. Nothing was removed. |
containers paused by neverwiped could not be resumed (…): run docker unpause …, then uninstall again. Nothing was removed. | A run was killed while Docker containers were paused for a volume copy, and they still cannot be resumed. | Run the docker unpause shown, check docker ps, then neverwiped uninstall again. Nothing was removed, so the list of paused containers is kept. |
Warning: the platform could not be notified: … | The platform couldn’t be told about the uninstall. | Delete the server in your dashboard if you want its backups gone. |
Warning: could not remove …: … | A file or folder of the agent could not be deleted (read-only file system, immutable attribute, busy mount). | Fix the cause (lsattr, mount), then neverwiped uninstall --yes again, or delete the path shown by hand. |
uninstall incomplete: … path(s) could not be removed (see the warnings above); fix the cause and run uninstall again | Some paths of the agent are still on the server (listed in the warnings above it). | Fix the cause and run neverwiped uninstall --yes again; it only removes what is left. |
unreadable list of backups: … | restic returned a list the agent can’t read. | Run neverwiped list again; if it persists, send neverwiped version to support. |
Commands and options
| Message, as printed | Cause | Fix |
|---|---|---|
this configuration belongs to the server … (ID …) and was copied to another machine (its /etc/machine-id differs): two machines must never share one server's keys. On a copy (cloned VM, copied image or folder): run “neverwiped uninstall --yes” here (it removes the copied keys from this machine only and leaves the other server untouched), then install again with the command of your client area: this machine becomes a new server with its own keys. If this machine IS that server and only its machine ID changed: run “neverwiped install --same-server” | This machine runs with the configuration of another server (cloned VM, copied image or /etc/neverwiped folder): it would share that server’s agent key, storage login and encryption key, so the agent refuses to back up, check or unregister here. | On the copy: neverwiped uninstall --yes (removes the copied keys here only; the other server is not told and keeps working), then the install command of your dashboard: the copy becomes a new server with its own keys and recovery key. On the original server whose machine ID changed: neverwiped install --same-server. |
usage: neverwiped check [--read 5%|1/26|2G] [--full] | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
invalid part to read: … (examples: 5%%, 1/26, 2G, all) | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
usage: neverwiped restore <date|id|latest> [--target FOLDER] [--include PATH] [--yes] [--plan | --approve CODE] | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
--max-seconds …: expected a number of seconds above 0 | --max-seconds is not a number of seconds above 0. | Give a number, for example --max-seconds 300, or leave it out for no limit. |
unknown kind … (bug, idea, help or other) | Unknown --kind. | bug, idea, help or other. |
neverwiped is not installed on this server: add --api-key KEY (account key, client area > API keys) | No agent key on this server and no account key given. | Add --api-key nw_acc_…. |
empty message: write it after the command (neverwiped feedback "…") or on the standard input | No message given. | neverwiped feedback "your message", or pipe it in. |
message too long (… characters, at most …) | The message is over the limit. | Shorten it, or write to support for long reports. |
nothing sent | You declined to send. | Nothing was sent. |
sending the message: … | The platform didn’t accept the message; the reason follows. | Try again later, or write to support by email. |
usage: neverwiped limits [--limit-upload RATE|off] [--limit-cpus N|off] [--priority low|normal] [--json] | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
unknown command … (see “neverwiped help”) | Unknown command. | Commands: install, status, backup now, list, restore, recover, feedback, uninstall, version, help. |
option --… needs a value | A flag is missing its value. | --target /path, --api-key nw_acc_…. |
unsupported language … (en, fr or es) | --lang or NEVERWIPED_LANG has an unknown value. | en, fr or es. |
unknown option … | Unknown flag. | See the command reference. |
neverwiped must run as root (for example: sudo neverwiped …) | Not root. | sudo neverwiped …. |
Note: --repo-key: the key is taken as a copy of the repository key of …, so the check code is not compared. restic checks it after the storage credential is re-issued; a wrong key is reported then. | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
Warning: the platform did not record the check code … of the new key: this server's recovery key was replaced too many times. Until support records it, “neverwiped recover” refuses the new key; the previous keys still work. Write to support, and keep your previous key as well. | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
--target and --include must not contain line breaks, escape sequences or invisible characters (bidi, zero-width, tags): …. Nothing was written. Check where this value came from before running anything | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
Warning: the platform already holds another check code for this server's recovery key (this server's code is …). Compare it with the code shown in your client area and with the paper where you wrote the key. | The platform already holds a different check code for this server’s recovery key; only this server’s agent key can set it, once. | Compare the code in neverwiped status, on your paper and on the server page; if they differ, write to support with neverwiped feedback. |
choose either --keep-backups or --delete-backups | Both --keep-backups and --delete-backups were given. | Pick one: neverwiped uninstall --yes --keep-backups or neverwiped uninstall --yes --delete-backups. |
this agent build has no release key: self-update is disabled. Update with the install script (see the command reference, “neverwiped update”) | This agent was built without a release key, so it cannot check a release and refuses to update itself. | Update with curl -fsSL https://neverwiped.com/install.sh | NEVERWIPED_NO_SETUP=1 sh: configuration and keys are kept. |
list of releases refused, nothing was installed: … | The list of releases is not signed by a key built into this agent, its signature is wrong, it has expired, or it is older than one this server already accepted. | Nothing was installed and backups go on. Do not work around it: write to support with neverwiped feedback, quoting the message. |
usage: neverwiped update [--check] [--to VERSION] [--yes] | --rollback | --pin VERSION|X.Y|none | --auto on|off (one at a time) | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
invalid pin …: give a version (0.3.1), a line (0.3) or none | --pin is neither a version, a line nor none. | Use for example --pin 0.3.1, --pin 0.3 or --pin none. |
no release to install: … | The version asked for (or pinned) is not published, was withdrawn, or has no binary for this architecture. | See what is available with neverwiped update --check; pin another version or --pin none. |
update cancelled, nothing was changed | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
update to … failed, the current version stays in place: … | The download failed, the binary did not match its signed SHA-256 or size, or it did not start and report the expected version. | The current version stays in place and backups go on. Try again later; if it repeats, write to support with neverwiped feedback. |
no previous version is kept on this server (no update was installed here yet): use neverwiped update --to VERSION | No update was installed on this server by neverwiped update, so there is no previous binary to put back. | Install a given version with neverwiped update --to VERSION. |
rollback failed, the current version stays in place: … | The kept binary could not be started, or could not be copied in place. | The current version stays. Install a given version with neverwiped update --to VERSION. |
unexpected answer from the platform: … | The platform answered with something that isn’t its API (proxy, captive portal, wrong --api-url). | Check --api-url / NEVERWIPED_API_URL and any outbound proxy. |
reading …: … | /etc/neverwiped/config.json can’t be read (permissions, damaged file). | Run as root; if the file is damaged, neverwiped recover on this server after moving /etc/neverwiped away. |
rate … is under 1 KiB/s (8 kbit/s), the smallest limit restic accepts | The upload limit is below 1 KiB/s, the smallest value restic takes. | Give at least 8kbit; a usable limit is rather a few Mbit/s. |
invalid rate …: give a number and a unit, for example 20mbit, 800kbit, 2MB or 500KiB (or off for no limit) | --limit-upload has no unit, or one the agent does not know. | Add a unit: 20mbit (or 20Mbps), 800kbit, 2MB, 500KiB; or off for no limit. |
invalid number of cores …: give a whole number from 1 to …, or off | --limit-cpus is not a whole number between 1 and the cores of this server. | Give for example --limit-cpus 1, or off to use every core (nproc shows how many there are). |
invalid priority …: low or normal | --priority is neither low nor normal. | Use --priority low (default) or --priority normal. |
no answer possible (non-interactive input): run again with --yes to confirm without a question | A question needs an answer and there is no terminal. | Add --yes; for a restore, prefer --target FOLDER. |
Answers from the platform (printed after a prefix such as “server registration:”)
| Message, as printed | Cause | Fix |
|---|---|---|
Too many requests, try again later. | Rate limit of the API. | Wait a little and try again. |
Invalid or revoked API key. Create a key in your client area, “API keys” section. | Key mistyped or revoked, email not confirmed yet, or account disabled. | Confirm your email (link sent at sign-up), then create a new key. |
This key cannot do this operation (account key nw_acc_… or agent key nw_agt_… expected). | An API token (nw_tok_…) was given to the agent. Tokens only work with the public API (/api/v1). | Use an account key (nw_acc_…) from the dashboard for install, recover and feedback; backups use the agent key written at install. |
This key cannot do this operation (account key nw_acc_… or agent key nw_agt_… expected). | A server key (nw_agt_…) was used instead of an account key. | Use an nw_acc_… key. |
This key cannot do this operation (account key nw_acc_… or agent key nw_agt_… expected). | The other kind of key was used: an account key (nw_acc_…) for install, recover and feedback, the agent key for backups. | Use an account key from the dashboard, API keys. |
This server was removed from your account. | The server was deleted in the dashboard, or uninstalled. | Reinstall: rm -f /etc/neverwiped/config.json, then neverwiped install --api-key … --yes (a new repository is created). |
Invalid JSON body. | The request was cut or altered on the way (proxy). | Run again; check any outbound proxy. |
Too many requests, try again later. | Too many registrations close together. | Try again later. |
No active subscription on this account. Choose a plan in your client area. | No subscription and no trial. | Choose a plan in the dashboard. |
Your subscription does not allow new backups (trial over or suspended). | Trial ended, payment overdue or cancellation. | Pay or choose a plan. |
Your plan's server limit is reached. Remove a server or move to a higher plan. | Your plan’s server limit is reached. | Delete an old server in the dashboard, or move up a plan. |
The storage is temporarily unavailable. Try again in a few minutes. | Storage under maintenance. | Try again later. |
New backups are suspended for this account (inactive subscription). Restore is still possible. | Subscription inactive. | Sort it out in the dashboard; restore still works. |
Unknown or already finished run. | The platform no longer knows this backup run (finished, or too old). | Nothing to do; the next backup reports normally. |
Unknown or already finished run. | Same, for another step of the run. | Nothing to do. |
Too many requests, try again later. | Too many pruning requests. | Nothing to do: retried next time. |
The cleanup of old backups is not due yet. | Internal or rare error; the text after the colon gives the detail. | Run neverwiped status and tail -n 50 /var/log/neverwiped.log; if it persists, send both to support. |
The cleanup of old backups is not due yet. | Old backups are cleaned at most once per period. | Nothing to do. |
The storage is temporarily unavailable. Try again in a few minutes. | Storage under maintenance. | Pruning is retried next time. |
Too many requests, try again later. | More than 5 new recovery keys for this server in an hour. | Wait an hour; the next backup sends the code of the key you already have. |
Too many replacements of this server's recovery key: write to support. | This server has replaced its recovery key 10 times; old codes are never dropped to make room. | The new key opens the backups, but recover refuses it until support records its code; keep the previous key too. Write to support. |
The check code of this server's recovery key is already recorded. | Same, as answered by the platform. | See the warning above. |
Too many requests, try again later. | Too many recoveries close together. | Try again later. |
Unknown server on this account. | --server names no server of this account. | Check the ID in the dashboard; recover without --server lists them. |
This server was removed from your account. | The server and its backups were deleted. | Nothing to recover. |
Too many requests, try again later. | Too many feedback messages close together. | Try again later. |
Still stuck?
Write to support from your account’s email with the output of neverwiped status and the last 50 lines of /var/log/neverwiped.log. Never send /etc/neverwiped/config.json, repo.key or your recovery key.