The old read -r line approach buffered until \n, but rsync --info=progress2
uses \r without \n for progress updates. Using tee(1) preserves \r
characters and writes through immediately for real-time progress.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
rsync --info=progress2 uses \r without \n, so the entire progress
stream is one long line. Split on both \r and \n to parse each
progress update independently.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add --info=progress2 --no-inc-recursive to rsync for overall progress
- Parse rsync progress output (percentage, speed, ETA) from log file
- Show ProgressBar widget and progress label below buttons
- Progress bar auto-hides when job finishes
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
View Log now displays the log in a RichLog panel below the buttons
instead of opening a modal screen. The log tails the file in real-time
with a 0.3s poll interval while the job is running.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The poll timer was reading from job.output in memory, which depended on
run_job's async task populating it. Now OperationLog reads new content
directly from the log file using seek, making it independent of the
async task. Also store the asyncio task reference to prevent GC.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Detects status from log content: Success (backup completed, no errors),
Failed (ERROR/FATAL markers), OK (lock released), Interrupted (no clean exit).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
asyncio's SubprocessTransport sends SIGKILL to child processes during
event loop cleanup, killing the gniza bash wrapper when the TUI exits.
Switch to subprocess.Popen which has no such cleanup behavior, allowing
backup jobs to continue running after the TUI is closed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The settings screen wrote WORK_DIR="/usr/local/gniza/workdir" (root-mode
path) into gniza.conf. load_config then overrode the correct user-mode
path set by detect_mode, causing mktemp and rsync log redirects to fail
with Permission denied — crashing the bash script while rsync continued
as an orphan.
- Remove WORK_DIR from AppSettings model and settings screen
- Re-run detect_mode after config load to restore correct paths
- Removed stale WORK_DIR from user's gniza.conf
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Save finished jobs to registry (24h TTL) so they survive TUI restart
- Fix PermissionError in PID check incorrectly marking alive processes as dead
- Handle CancelledError explicitly to preserve running status on TUI exit
- Tail log files for reconnected running jobs instead of showing stale output
- Detect actual return code from log content; show "?" for unknown status
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The subprocess stdout was a pipe to the TUI. When the TUI exited, the
pipe broke (SIGPIPE) and killed the backup process. Now the subprocess
writes to a log file in WORK_DIR, and the TUI tails it for live
display. When the TUI exits, the subprocess keeps running because it
writes to a file, not a pipe. On restart, the log file is loaded to
show output for reconnected or finished background jobs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Previously _load_registry silently dropped dead PIDs, so jobs that
completed in the background were invisible on restart. Now dead PIDs
are loaded as 'success' status so the user sees they completed. The
registry is cleaned up after loading.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Previously action_quit() killed all running jobs before exiting,
making it impossible to reconnect to them on restart. Now shows a
confirmation dialog when jobs are running and lets them continue
in the background.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace tee with _snaplog_tee shell function that writes each line
unbuffered to three destinations: snapshot log, application log file,
and stderr (which the TUI captures for live display). This fixes the
issue where rsync file-by-file output was invisible in both the Logs
screen and the Running Tasks view due to pipe buffering.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Jobs are now saved to gniza-jobs.json in WORK_DIR when they start and
finish. On TUI restart, the registry is loaded and PIDs are checked —
still-running jobs appear in the Running Tasks screen and can be
killed. Reconnected jobs are polled every second to detect completion.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When TARGET_INCLUDE has a directory pattern like 'embajada/', rsync
would include the directory but exclude its contents (files didn't
match any include rule before hitting --exclude=*). Now adds a
'pattern**' include for each directory pattern.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Save log, rsync_error, summary, and index files into each snapshot
directory after backup. Rsync/rclone output is captured via tee during
transfer so the TUI still shows live output while logging to disk.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Jobs were tied to screen @work tasks, so switch_screen cancelled the
worker and lost the process reference. Now start_job uses
asyncio.create_task so jobs survive screen changes and can be killed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
SIGTERM was being ignored by child processes. Use SIGKILL via the
actual process group ID, with proc.kill() as fallback.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Start CLI subprocesses in their own session so SIGTERM via
os.killpg reaches child processes (rsync, etc.) not just the shell.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Backup and restore operations now run as background jobs instead of
blocking modal screens. Users can navigate away and check progress
from a dedicated Running Tasks screen. OperationLog supports attaching
to running jobs with live output polling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Exit 23 (partial transfer, permission denied) and 24 (vanished files)
are expected in non-root backups. These no longer trigger retries or
fail the backup — they log a warning and continue.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Utility commands (schedule, snapshots, remotes, etc.) were creating
empty 0 B log files on every invocation. Now init_logging only runs
for commands that produce meaningful log output.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Cron runs with a very limited PATH that may not include sshpass, rsync,
etc. Add explicit PATH to generated cron lines and log output to
cron.log instead of /dev/null for debugging.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
grep returns exit 1 when no match, which with set -eo pipefail kills
the script. Add || true to the grep pipeline in credential extraction.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The grep -v pipeline combined with set -eo pipefail caused the entire
installer to exit silently when the user answered "y" to web dashboard.
Replace with proper if/else and add user-level systemd service support
(systemctl --user) for non-root installs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Apply shquote() to all remaining remote_exec paths that interpolate
variables into single-quoted shell strings. Covers list/resolve/clean
snapshots, symlink updates, retention pruning, and restore file listing.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add shquote() to escape single quotes in paths passed to remote_exec,
preventing shell injection via REMOTE_BASE containing single quotes
- Apply shquote to remote_exec calls in remotes.sh, backup.sh, transfer.sh, ssh.sh
- Add DISK_USAGE_THRESHOLD validation in config.sh
- Export SMTP_PASSWORD (was missing from export list)
- Fix WEB_PORT default mismatch: use 2323 consistently in from_conf and settings save
- Narrow exception catch in remotes.py disk info fetch to KeyError/LookupError
- Quote REMOTE_KEY in build_rsync_ssh_cmd for paths with spaces
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Parse df output locally using awk to find the % field and extract
the 3 size fields before it. Handles wrapped lines and CR chars.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Move awk parsing to the remote to avoid SSH output encoding issues
that caused empty fields when parsing locally.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
df --output=pcent is not available on all systems. Use grep -oP to
extract the percentage field from standard df output.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use df --output=pcent for reliable single-column output instead of
parsing multi-column df output through awk over SSH.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Disk usage threshold (default 95%) can now be controlled from
Settings. Set to 0 to disable the check.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add TARGET_INCLUDE field for rsync include patterns (comma-separated)
- Pass TARGET_INCLUDE and TARGET_EXCLUDE to rsync in transfer_folder
- Include mode uses --include='*/' + patterns + --exclude='*' + --prune-empty-dirs
- Abort backup if remote disk usage >= 95%
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Disk info is now shown directly in the table. Speed test and detailed
disk info buttons/commands are no longer needed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Disk Info: runs df -h and df -i on remote via SSH (or locally)
- Speed Test: uploads/downloads 10MB test file via rsync, measures Mbps
- Both available as CLI commands: gniza remotes disk-info/speed-test --name=NAME
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
