Make sure to populate the cache in cache_space_refresh() at least once
otherwise it's possible that the system boots fast enough (and the journal
flush service is finished) before the invalidate cache timeout (30 us) has
expired.
Fixes: #4790
Let's remove chase_symlinks_prefix() and instead introduce a flags parameter to
chase_symlinks(), with a flag CHASE_PREFIX_ROOT that exposes the behaviour of
chase_symlinks_prefix().
Let's use chase_symlinks() everywhere, and stop using GNU
canonicalize_file_name() everywhere. For most cases this should not change
behaviour, however increase exposure of our function to get better tested. Most
importantly in a few cases (most notably nspawn) it can take the correct root
directory into account when chasing symlinks.
We don't have plural in the name of any other -util files and this
inconsistency trips me up every time I try to type this file name
from memory. "formats-util" is even hard to pronounce.
I'm seeing strange decompression errors with lz4, which
might be content-dependent. Extend test-compression to allow
testing specific content.
(Edit: PEBKAC: lzcat and lz4cat are not the same beast.
Nevertheless, the test might still be useful in the future.)
This stripping is contolled by a new boolean parameter. When the parameter
is true, it means that the caller does not care about the distinction between
initrd and real root, and wants to act on both rd-dot-prefixed and unprefixed
parameters in the initramfs, and only on the unprefixed parameters in real
root. If the parameter is false, behaviour is the same as before.
Changes by caller:
log.c (systemd.log_*): changed to accept rd-dot-prefix params
pid1: no change, custom logic
cryptsetup-generator: no change, still accepts rd-dot-prefix params
debug-generator: no change, does not accept rd-dot-prefix params
fsck: changed to accept rd-dot-prefix params
fstab-generator: no change, custom logic
gpt-auto-generator: no change, custom logic
hibernate-resume-generator: no change, does not accept rd-dot-prefix params
journald: changed to accept rd-dot-prefix params
modules-load: no change, still accepts rd-dot-prefix params
quote-check: no change, does not accept rd-dot-prefix params
udevd: no change, still accepts rd-dot-prefix params
I added support for "rd." params in the three cases where I think it's
useful: logging, fsck options, journald forwarding options.
This makes journald use the common option parsing functionality.
One behavioural change is implemented:
"systemd.journald.forward_to_syslog" is now equivalent to
"systemd.journald.forward_to_syslog=1".
I think it's nicer to use this way.
Now that determine_space_for() only deals with storage space (cached) values,
rename it so it reflects the fact that only the cached storage space values are
updated.
Updating min_use is rather an unusual operation that is limited when we first
open the journal files, therefore extracts it from determine_space_for() and
create a function of its own and call this new function when needed.
determine_space_for() is now dealing with storage space (cached) values only.
There should be no functional changes.
The set of storage space values we cache are calculated according to a couple
of filesystem statistics (free blocks, block size).
This patch caches the vfs stats we're interested in so these values are
available later and coherent with the rest of the space cached values.
This patch makes system_journal_open() stop emitting the space usage
message. The caller is now free to emit this message when appropriate.
When restarting the journal, we can now emit the message *after*
flushing the journal (if required) so that all flushed log entries are
written in the persistent journal *before* the status message.
This is required since the status message is always younger than the
flushed entries.
Fixes#4190.
This commit simply extracts from determine_space_for() the code which emits the
storage usage message and put it into a function of its own so it can be reused
by others paths later.
No functional changes.
This structure keeps track of specificities for a given journal type
(persistent or volatile) such as metrics, name, etc...
The cached space values are now moved in this structure so that each
journal has its own set of cached values.
Previously only one set existed and we didn't know if the cached
values were for the runtime journal or the persistent one.
When doing:
determine_space_for(s, runtime_metrics, ...);
determine_space_for(s, system_metrics, ...);
the second call returned the cached values for the runtime metrics.
This commit simply extracts from determine_space_for() the code which
determines the FS usage where the passed path lives (statvfs(3)) and put it
into a function of its own so it can be reused by others paths later.
No functional changes.
Never permit that we write to journal files that have newer timestamps than our
local wallclock has. If we'd accept that, then the entries in the file might
end up not being ordered strictly.
Let's refuse this with ETXTBSY, and then immediately rotate to use a new file,
so that each file remains strictly ordered also be wallclock internally.
As soon as we notice that the clock jumps backwards, rotate journal files. This
is beneficial, as this makes sure that the entries in journal files remain
strictly ordered internally, and thus the bisection algorithm applied on it is
not confused.
This should help avoiding borked wallclock-based bisection on journal files as
witnessed in #4278.
Let's use the earliest linearized event timestamp for journal entries we have:
the event dispatch timestamp from the event loop, instead of requerying the
timestamp at the time of writing.
This makes the time a bit more accurate, allows us to query the kernel time one
time less per event loop, and also makes sure we always use the same timestamp
for both attempts to write an entry to a journal file.
When iterating through partially synced journal files we need to be prepared
for hitting with invalid entries (specifically: non-initialized). Instead of
generated an error and giving up, let's simply try to preceed with the next one
that is valid (and debug log about this).
This reworks the logic introduced with caeab8f626
to iteration in both directions, and tries to look for valid entries located
after the invalid one. It also extends the behaviour to both iterating through
the global entry array and per-data object entry arrays.
Fixes: #4088
Let's make dissecting of borked journal files more expressive: if we encounter
an object whose first 8 bytes are all zeroes, then let's assume the object was
simply never initialized, and say so.
Previously, this would be detected as "overly short object", which is true too
in a away, but it's a lot more helpful printing different debug options for the
case where the size is not initialized at all and where the size is initialized
to some bogus value.
No function behaviour change, only a different log messages for both cases.
Let's make it easier to figure out when we see an invalid journal file, why we
consider it invalid, and add some minimal debug logging for it.
This log output is normally not seen (after all, this all is library code),
unless debug logging is exlicitly turned on.
When date is changed in system to future and normal user logs to new journal file, and then date is changed back to present time, the "journalctl --list-boot" command goes to forever loop. This commit tries to fix this problem by checking first the boot id list if the found boot id was already in that list. If it is found, then stopping the boot id find loop.
This adds a new invocation ID concept to the service manager. The invocation ID
identifies each runtime cycle of a unit uniquely. A new randomized 128bit ID is
generated each time a unit moves from and inactive to an activating or active
state.
The primary usecase for this concept is to connect the runtime data PID 1
maintains about a service with the offline data the journal stores about it.
Previously we'd use the unit name plus start/stop times, which however is
highly racy since the journal will generally process log data after the service
already ended.
The "invocation ID" kinda matches the "boot ID" concept of the Linux kernel,
except that it applies to an individual unit instead of the whole system.
The invocation ID is passed to the activated processes as environment variable.
It is additionally stored as extended attribute on the cgroup of the unit. The
latter is used by journald to automatically retrieve it for each log logged
message and attach it to the log entry. The environment variable is very easily
accessible, even for unprivileged services. OTOH the extended attribute is only
accessible to privileged processes (this is because cgroupfs only supports the
"trusted." xattr namespace, not "user."). The environment variable may be
altered by services, the extended attribute may not be, hence is the better
choice for the journal.
Note that reading the invocation ID off the extended attribute from journald is
racy, similar to the way reading the unit name for a logging process is.
This patch adds APIs to read the invocation ID to sd-id128:
sd_id128_get_invocation() may be used in a similar fashion to
sd_id128_get_boot().
PID1's own logging is updated to always include the invocation ID when it logs
information about a unit.
A new bus call GetUnitByInvocationID() is added that allows retrieving a bus
path to a unit by its invocation ID. The bus path is built using the invocation
ID, thus providing a path for referring to a unit that is valid only for the
current runtime cycleof it.
Outlook for the future: should the kernel eventually allow passing of cgroup
information along AF_UNIX/SOCK_DGRAM messages via a unique cgroup id, then we
can alter the invocation ID to be generated as hash from that rather than
entirely randomly. This way we can derive the invocation race-freely from the
messages.
Currently, the ratelimit does not handle the number of suppressed messages accurately.
Even though the number of messages reaches the limit, it still allows to add one extra messages to journal.
This patch fixes the problem.
This patch fixes wrong calculation of burst_modulate(), which now calculates
the values smaller than really expected ones if available disk space is
strictly more than 1MB.
In particular, if available disk space is strictly more than 1MB and strictly
less than 16MB, the resulted value becomes smaller than its original one.
>>> (math.log2(1*1024**2)-16) / 4
1.0
>>> (math.log2(16*1024**2)-16) / 4
2.0
>>> (math.log2(256*1024**2)-16) / 4
3.0
→ This matches the comment in the function.
Since commit 5996c7c295 (v190 !), the
calculation of the HMAC is broken because the hash for a data object
including a field is done in the wrong order: the field object is
hashed before the data object is.
However during verification, the hash is done in the opposite order as
objects are scanned sequentially.
Zbigniew Jędrzejewski-Szmek
Lennart Poettering
Franck Bui
Waldemar Brodkorb
Thomas Hindoe Paaboel Andersen
Umut Tezduyar Lindskog
hese10
Yuki Inoguchi
Vito Caputo
HATAYAMA Daisuke
Felix Zhang
Martin Pitt