| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This essentially reverts commit 66cdd5bc99a5 which aimed to make the
dss log facility C89 conform. While this was a worthwhile goal in
2012, it has little value today, since in 2025 we can safely assume
a C99 compliant compiler.
The patch was created with
git revert -Xours 66cdd5bc
followed by manual tweaks to make it compile again.
|
| |
|
|
|
|
| |
The doxygen comments actually make the source code harder to read,
because they tend to document the obvious. Besides, they are not used
uniformly accross the code base anyway.
|
| |
|
|
|
| |
It's sometimes handy to just type e.g. "dss help run" to see the
synopsis and the options of the ls subcommand.
|
| |
|
|
|
|
|
| |
This gets rid of Makefile.deps in favor of one auto-generated .d
file per .c file, and adds version-gen.sh, a shell script executed
by make(1) which tries to determine the version number that is going
to be incorporated into the executable via the generated version.c.
|
| |
|
|
|
| |
It's easier if dss.c passes the hash of the the config file path to
the ipc functions than to pass a char pointer.
|
| |
|
|
|
|
|
|
| |
Currently we call get_config_file_name() before each call call to a
public function of ipc.c because those functions compute the IPC key
from the filename. This is unnecessary and caused several bugs. Clean
up this mess by computing the filename only once and store it in a
global variable.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The dss lock works by first turning the given config file path
argument into a canonical absolute path using dss_realpath(), then
hashing this absolute path to obtain a key ID for semget(2).
If the given path is relative, we have to compute the ID before
changing to the destination directory because dss_realpath() needs to
call stat(2) to detect symlinks, and this system call will fail if the
current working directory has changed. This is currently not the case
as we change to the destination directory early in check_config().
If dss_realpath() fails, we silently use the unmodified path argument
for hashing to deal with the case that the default config does not
exist. As a result, if relative paths are given, the key ID depends
on whether or not change_to_dest_dir() was called. This is the case
for the run subcommanmd, but not for the kill subcommand. Thus the
kill subcommand does not work as expected if a relative path is given.
Fix this by grabbing the lock before changing the working directory
in all cases.
|
| |
|
|
|
|
| |
If parse_config_file() encounters an error, it logs the error *and*
returns the error code, which causes the error message to be logged
again. Fix this by removing the log statement from parse_config_file().
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A couple of cleanups and unifications for the snapshot pruning code,
which is executed by the prune and run subcommands. With the patches
applied, both subcommands behave identically, i.e. they consider
the same set of snapshots as candidates for pruning. Also the prune
command gained the new --disk-space option to force it to act as if
disk space was high or low.
Cooking for six weeks.
* refs/heads/t/prune:
Introduce prune --disk-space.
Revamp com_prune().
Factor out find_removable_snapshot().
prune Simplify rm exit code logic.
find_oldest_removable_snapshot(): Improve log message.
prune: Print a message if there is nothing to prune.
get_snapshot_list(): Add comment about sorting.
prune: Fail gracefully if pre-rm hook fails.
|
| | |
| |
| |
| |
| |
| |
| | |
The new option takes a mandatory argument of the set {check, high,
low}. The latter two make the command act as if disk space was
high/low without actually performing the check. This is mostly useful
for debugging.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The prune subcommand implements its own logic for picking the snapshot
to remove. The algorithm is similar but not identical to how the run
subcommand gets rid of snapshots.
This patch eliminates this inconsistency by changing com_prune()
to call the find_removable_snapshot() helper which was introduced in
the previous commit.
Since find_removable_snapshot() returns a dynamically allocated string
via the "why" pointer, we have to introduce another label for freeing
this memory.
The patch also improves the help text of the prune subcommand slightly.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Both the run and the prune subcommand contain code for picking a
suitable snapshot to remove, with slightly different semantics.
This is a preparatory patch for eliminating the differences between
the two implementations. It introduces a new helper which can be
shared. After this patch, only the run subcommand uses the helper. A
subsequent patch will convert the second caller, com_prune().
|
| | |
| |
| |
| |
| | |
We can rely on the fact that wait_for_remove_process() sets
snapshot_removal_status to the correct value on success.
|
| | |
| |
| |
| |
| |
| | |
The other three functions which return a candidate for removal print
one debug message but not the name of the selected snapshot, so make
find_oldest_removable_snapshot() follow this pattern too.
|
| | |
| |
| |
| |
| |
| | |
When the prune subcommand finds there are no snapshots to prune, the
command stays silent, which is confusing. Make it print "nothing to
prune" in this case.
|
| | |
| |
| |
| |
| |
| | |
In this case wait_for_remove_process() returns non-negative and we
miss to set the exit code, making the command appear to succeed even
if the rm process has not been created.
|
| | |
| |
| |
| |
| |
| | |
This bug was introduced recently in commit dd3f58c0bdf1
(create_rsync_argv(): Allocate correctly sized arg array) which
redefined N.
|
| |/
|
|
|
| |
Both callers pass the same value for the delimiters, so we can
remove the last argument of the function.
|
| |\
| |
| |
| |
| |
| |
| |
| | |
Was cooking for more than a year.
* refs/heads/t/multiple-source-dirs:
create_rsync_argv(): Allocate correctly sized arg array.
Support multiple source directories.
|
| | |
| |
| |
| |
| |
| | |
In the calculation of the length of the argv array we did not take
into account that --source-dir may be given multiple times. This can
result in an invalid write at the end of the allocated space.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
rsync is capable of copying multiple source directories to a single
destination, but this is currently not supported by dss. This commit
adds this functionality. The implementation is straight-forward,
except that we don't want to add a trailing slash to every source
directory. The new comment in dss.c explains this in more detail.
Suggested-By: Sanja Jasek <sanja.jasek@tuebingen.mpg.de>
Tested-By: Sanja Jasek <sanja.jasek@tuebingen.mpg.de>
|
| | |
| |
| |
| |
| |
| | |
This comment went stale ten years ago in commit 360bcc95d588 (Clean
up snapshot removal logic) which changed the type of the return value
of find_redundant_snapshot() from int to struct snapshot *.
|
| |/
|
|
| |
The old comment did not provide any information at all..
|
| |
|
|
| |
Signed-off-by: Andre Noll <maan@tuebingen.mpg.de>
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When handle_signal(), the signal dispatcher of the run subcommand,
detects that SIGINT or SIGTERM was received, it calls kill_children()
to terminate any running rsync or rm processes. It then returns
negative which terminates the select loop. However, after select_loop()
returns, kill_children() is called again. Also the error message is
logged twice.
Not a biggie, but let's get rid of this redundancy by removing the
first call to kill_children().
Since handle_signal() is only called from com_run(), this patch
affects only the run subcommand.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The function computes the average idle time between snapshots and adds
this value to the completion time of the last snapshot to obtain the
start time for the next snapshot.
However, if the last snapshot happens to be incomplete, its completion
time is set to -1. Hence the computed next snapshot time is going to
be in the past, so we start the next snapshot immediately.
Although this is incorrect, the bug is benign because the correct next
snapshot time should also be in the past since we decided earlier to
create the snapshot which was now found incomplete.
Fix this by using the completion time of the last _complete_ snapshot
instead.
|
| |
|
|
|
|
|
| |
We never pass a NULL pointer to create_snapshot(), but scan-build
is unable to prove this and claims that the array access results in
a null pointer dereference. The added assertion helps the reader of
the code, and it quietens scan-build.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This gets rid of existing copyright templates in favor of just the
one-liner SPDX (Software Package Data Exchange) notice. All files are
licensed under the GPL-2.0, so the same tag is added to each file. No
copyright is changed by this commit.
Several files (mostly the very short ones) did not contain a license
text so far. By default all files without license information are
under the default license of this package, which is GPL version 2.
This commit adds the missing SPDX line so that now all files except
dss.css, index.html.in, INSTALL, NEWS and README have it.
We also remove author and copyright year, since the author is the same
everywhere, and the year hasn't been updated any more since at least
six years. Accurate information is available from the git log.
The COPYING file can also be removed because the license text at
https://spdx.org/licenses/GPL-2.0.html is immutable.
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Two patches which make life easier for shutdown scripts which need
to terminate the dss process, but would like to wait until the exit
hook completed.
The merge conflicted in dss.suite, but this was trivial to fix.
Cooking for a week.
* refs/heads/t/kill-w:
kill: New option --wait.
run: Wait for children to die.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Simply running "dss kill" during system shutdown to terminate the
dss process does not work as expected because the kill subcommand
exits after the signal has been sent, which might be long before the
targeted dss process terminates.
For example, the dss main process might be running its exit hook to
inform the system administrator about the fact that the dss service
is going down when the shutdown procedure already has deactivated the
network. Or the shutdown procedure kills the exit hook with SIGKILL
during its normal "killing remaining processes" phase before file
systems are unmounted.
With the --wait option, the kill subcommand will not return until the
dss process has died or the timeout expires. We hardcode the timeout
in send_signal() for the time being. It can be made configurable if
this turns out to be necessary.
|
| | |
| |
| |
| |
| |
| |
| |
| | |
When the select loop returns and dss is about to terminate, it sends
SIGTERM to any running rm or rsync processes and exits. It does not
wait for these processes to die, however. This is trivial to implement,
and it makes life easier for shutdown scripts which like to proceed
with unmounting file systems.
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A fix for a long standing issue with the exit hook.
Cooking for a week.
* refs/heads/t/exit-hook:
Allow word-splitting for exit hook.
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
All hooks except the exit hook are run via dss_exec_cmdline_pid(),
which performs word splitting to create the argument vector for
exec(2). For the exit hook, however, we build the argument vector
manually, so the command line for the exit hook is not split.
This commit removes this inconsistency. However, we can't use
dss_exec_cmdline_pid() here because we need to append the error string
which caused dss to exit to the argument vector as a single argument,
and this string may well contain whitespace characters.
Hence we run split_args() on the argument to --exit-hook to obtain
an argument vector, append the error string as another element,
and then run dss_exec().
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A simple patch which improves the ls output.
Cooking for 10 days.
* refs/heads/t/ls-incomplete-duration:
ls: Print current duration of incomplete snapshots.
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| | |
Currently the duration of incomplete (and orphaned) snapshots is shown
as 0:00. It's more interesting to see for how long the snapshot is
already being created, so print the difference of the current time
and the start time instead.
Fix an overlong line and a whitespace issue while at it.
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A new main option which aborts dss if the destination file system is
not mounted.
Cooking for two weeks.
* refs/heads/t/mountpoint:
New option: --mountpoint.
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| | |
The new option applies to run, create, ls and prune. The feature
could be implemented as a pre-create hook, but since it is so common,
it makes sense to add it to dss proper.
As for the implementation we simply check that "." and ".." are on
different devices (or are identical).
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A few patches which change dss_log() to only print the time stamp
and the function if dss is executing the "run" subcommand.
Cooking for a week.
* refs/heads/t/short-log:
Subcommand sensitive logging.
Save the subcommand pointer in a global variable.
main(): Don't initialize user data pointer.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It's kind of silly to log the current time and the function name for
subcommands other than "run", since these commands run only for a
short time.
This commit changes dss_log() to only print this additional information
when dss run is being executed.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is needed for subcommand sensitive logging which will be
introduced in a subsequent commit. For now it allows to drop the
argument of check_config(), which is good given that handle_sighup()
already played dirty games by "knowing" it is only called from the
run subcommand.
|
| | |/
| |
| |
| |
| |
| | |
It is never used uninitialized. The unnecessary initialization only
hides the warning from the compiler if it ever should ever be used
uninitialized.
|
| | |
| |
| |
| |
| |
| | |
It is kind of obvious that find_orphaned_snapshot() looks for, well,
orphaned snapshots. The new message at least gives the user a vague
idea what this means.
|
| | |
| |
| |
| | |
Sending a signal is a significant event which deserves a log message.
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A single patch which adds the new configtest subcommand, plus a fixup
for a formatting issue which was noticed only after the branch had
already been merged to next.
* refs/heads/t/configtest:
show_subcommand_summary(): Increase column width.
New subcommand: configtest.
|
| | | |
| | |
| | |
| | |
| | | |
Since the new configtest command is 10 characters long, there was no
whitespace between the command name and the short description.
|
| | |/
| |
| |
| |
| |
| |
| |
| | |
Similar to the identically named subcommand of apache2ctl. This is
trivial to implement because we only need to describe the subcommand
in dss.suite and create a command handler which prints an OK message
and returns success. If the config file contains errors, we abort
earlier anyway.
|
| |\ \
| |/
|/|
| |
| | |
* maint:
Fix snapshot removal reason.
|
| | |
| |
| |
| |
| | |
If the oldest snapshot has to be removed because disk space is low,
we currently log "orphaned" as the reason, which is incorrect.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
It is considered good practice to run rsync with --checksum from time
to time. This patch implements the feature via the new --checksum
option.
The probabilistic approach was chosen so that dss does not need to
remember which snapshots were created with --checksum.
|
