| Age | Commit message (Collapse) | Author |
|---|
|
Due to the substantial interface changes in cgroup unified hierarchy, new IO
settings are introduced. Currently, IO settings apply only to unified
hierarchy and BlockIO to legacy. While the transition is necessary, it's
painful for users to have to provide configs for both. This patch implements
translation from one config set to another for configs which make sense.
* The translation takes place during application of the configs. Users won't
see IO or BlockIO settings appearing without being explicitly created.
* The translation takes place only if there is no config for the matching
cgroup hierarchy type at all.
While this doesn't provide comprehensive compatibility, it should considerably
ease transition to the new IO settings which are a superset of BlockIO
settings.
v2:
- Update test-cgroup-mask.c so that it accounts for the fact that
CGROUP_MASK_IO and CGROUP_MASK_BLKIO move together. Also, test/parent.slice
now sets IOWeight instead of BlockIOWeight.
|
|
Factor out the following functions out of cgroup_context_apply()
* cgroup_context_[blk]io_weight()
* cgroup_apply_[blk]io_device_weight()
* cgroup_apply_[blk]io_device_limit()
This is pure refactoring and shouldn't cause any functional differences.
|
|
CGroupBlockIODeviceBandwith is used to keep track of IO bandwidth limits for
legacy cgroup hierarchies. Unlike the unified hierarchy counterpart
CGroupIODeviceLimit, a CGroupBlockIODeviceBandwiddth records either a read or
write limit and has a couple issues.
* There's no way to clear specific config entry.
* When configs are cleared for an IO direction of a unit, the kernel settings
aren't cleared accordingly creating discrepancies.
This patch updates CGroupBlockIODeviceBandwidth so that it behaves similarly to
CGroupIODeviceLimit - each entry records both rbps and wbps limits and is
cleared if both are at default values after kernel settings are updated.
|
|
cgroup IO controller supports maximum limits for both bandwidth and IOPS but
systemd resource control currently only supports bandwidth limits. This patch
adds support for IOReadIOPSMax and IOWriteIOPSMax when unified cgroup hierarchy
is in use.
It isn't difficult to also add BlockIOReadIOPS and BlockIOWriteIOPS for legacy
hierarchies but IO control on legacy hierarchies is half-broken anyway, so
let's leave it alone for now.
|
|
Currently, there are two cgroup IO limits, bandwidth max for read and write,
and they are hard-coded in various places. This is fine for two limits but IO
is expected to grow more limits - low, high and max limits for bandwidth and
IOPS - and hard-coding each limit won't make sense.
This patch replaces hard-coded limits with an array indexed by
CGroupIOLimitType and accompanying string and default value tables so that new
limits can be added trivially.
|
|
Makes it consistent with the other branches here.
|
|
core: add io controller support on the unified hierarchy
|
|
We currently generate log message about unit being started even when
unit was started already and job didn't do anything. This is because job
was requested explicitly and hence became anchor job of the transaction
thus we could not eliminate it. That is fine but, let's not pollute
journal with useless log messages.
$ systemctl start systemd-resolved
$ systemctl start systemd-resolved
$ systemctl start systemd-resolved
Current state:
$ journalctl -u systemd-resolved | grep Started
May 05 15:31:42 rawhide systemd[1]: Started Network Name Resolution.
May 05 15:31:59 rawhide systemd[1]: Started Network Name Resolution.
May 05 15:32:01 rawhide systemd[1]: Started Network Name Resolution.
After patch applied:
$ journalctl -u systemd-resolved | grep Started
May 05 16:42:12 rawhide systemd[1]: Started Network Name Resolution.
Fixes #1723
|
|
Private /dev will not be managed by udev or others, so we can make it
noexec and readonly after we have made all device nodes. As /dev/shm
needs to be writable, we can't use bind_remount_recursive().
|
|
unit_set_slice() fails with -EBUSY if the unit already has a slice associated
with it. This makes it impossible to override slice through dropin config or
over dbus. There's no reason to disallow slice changes as long as cgroups
aren't realized. Fix it.
Fixes #3240.
Signed-off-by: Tejun Heo <htejun@fb.com>
Reported-by: Davide Cavalca <dcavalca@fb.com>
|
|
Drop all dangling old /dev mounts before mounting a new private /dev tree.
|
|
This new method returns information by unit names. Instead of ListUnitsByPatterns
this method returns information of inactive and even unexisting units.
Moved dbus unit reply logic into a separate shared function.
Resolves https://github.com/coreos/fleet/pull/1418
|
|
Fix "preset-all" with dangling symlinks and install-section hint emitted too eagerly
|
|
|
|
don't reopen socket fds when reloading the daemon
|
|
|
|
Cgroup fixes.
|
|
core: use an AF_UNIX/SOCK_DGRAM socket for cgroup agent notification
|
|
|
|
Before:
$ systemctl show --property TriggerLimitIntervalSec test.socket
TriggerLimitIntervalSec=2000000
After:
$ systemctl show --property TriggerLimitIntervalUSec test.socket
TriggerLimitIntervalUSec=2s
|
|
Fixes: #3194
|
|
Previously, we'd simply close and reopen the socket file descriptors. This is
problematic however, as we won't transition through the SOCKET_CHOWN state
then, and thus the file ownership won't be correct for the sockets.
Rework the flushing logic, and actually read any queued data from the sockets
for flushing, and accept any queued messages and disconnect them.
|
|
Previously, when the daemon was reloaded and the configuration of a socket unit
file was changed so that a different set of socket ports was defined for the
socket we'd simply reopen the socket fds not yet open. This is problematic
however, as this means the SOCKET_CHOWN state is not run for them, and thus
their UID/GID is not corrected.
With this change, don't open the missing file descriptors, but log about this
issue, and ask the user to restart the socket explicit, to make sure all
missing fds are opened.
Fixes: #3171
|
|
This should bring no behavioural change.
|
|
On the unified hierarchy, blkio controller is renamed to io and the interface
is changed significantly.
* blkio.weight and blkio.weight_device are consolidated into io.weight which
uses the standardized weight range [1, 10000] with 100 as the default value.
* blkio.throttle.{read|write}_{bps|iops}_device are consolidated into io.max.
Expansion of throttling features is being worked on to support
work-conserving absolute limits (io.low and io.high).
* All stats are consolidated into io.stats.
This patchset adds support for the new interface. As the interface has been
revamped and new features are expected to be added, it seems best to treat it
as a separate controller rather than trying to expand the blkio settings
although we might add automatic translation if only blkio settings are
specified.
* io.weight handling is mostly identical to blkio.weight[_device] handling
except that the weight range is different.
* Both read and write bandwidth settings are consolidated into
CGroupIODeviceLimit which describes all limits applicable to the device.
This makes it less painful to add new limits.
* "max" can be used to specify the maximum limit which is equivalent to no
config for max limits and treated as such. If a given CGroupIODeviceLimit
doesn't contain any non-default configs, the config struct is discarded once
the no limit config is applied to cgroup.
* lookup_blkio_device() is renamed to lookup_block_device().
Signed-off-by: Tejun Heo <htejun@fb.com>
|
|
The unit file settings are called SocketUser= and SocketGroup= hence name these
fields that way in the "systemd-analyze dump" output too.
https://github.com/systemd/systemd/issues/3171#issuecomment-216216995
|
|
Let's lower the default values a bit, and pick different defaults for
Accept=yes and Accept=no sockets.
Fixes: #3167
|
|
The macro determines the right length of a AF_UNIX "struct sockaddr_un" to pass to
connect() or bind(). It automatically figures out if the socket refers to an
abstract namespace socket, or a socket in the file system, and properly handles
the full length of the path field.
This macro is not only safer, but also simpler to use, than the usual
offsetof() + strlen() logic.
|
|
dbus-daemon currently uses a backlog of 30 on its D-bus system bus socket. On
overloaded systems this means that only 30 connections may be queued without
dbus-daemon processing them before further connection attempts fail. Our
cgroups-agent binary so far used D-Bus for its messaging, and hitting this
limit hence may result in us losing cgroup empty messages.
This patch adds a seperate cgroup agent socket of type AF_UNIX/SOCK_DGRAM.
Since sockets of these types need no connection set up, no listen() backlog
applies. Our cgroup-agent binary will hence simply block as long as it can't
enqueue its datagram message, so that we won't lose cgroup empty messages as
likely anymore.
This also rearranges the ordering of the processing of SIGCHLD signals, service
notification messages (sd_notify()...) and the two types of cgroup
notifications (inotify for the unified hierarchy support, and agent for the
classic hierarchy support). We now always process events for these in the
following order:
1. service notification messages (SD_EVENT_PRIORITY_NORMAL-7)
2. SIGCHLD signals (SD_EVENT_PRIORITY_NORMAL-6)
3. cgroup inotify and cgroup agent (SD_EVENT_PRIORITY_NORMAL-5)
This is because when receiving SIGCHLD we invalidate PID information, which we
need to process the service notification messages which are bound to PIDs.
Hence the order between the first two items. And we want to process SIGCHLD
metadata to detect whether a service is gone, before using cgroup
notifications, to decide when a service is gone, since the former carries more
useful metadata.
Related to this:
https://bugs.freedesktop.org/show_bug.cgi?id=95264
https://github.com/systemd/systemd/issues/1961
|
|
make virtualization detection quieter, rework unit start limit logic, detect unit file drop-in changes correctly, fix autofs state propagation
|
|
|
|
This reverts commit 6d10d308c6cd16528ef58fa4f5822aef936862d3.
It got squashed by mistake.
|
|
Refuse Alias, DefaultInstance, templated units in install (as appropriate)
|
|
that way we can be sure that there's no expiry timeout in place at any time
when we aren't in the RUNNING state.
|
|
Port the progagation logic to the generic Unit->trigger_notify() callback logic
in the unit vtable, that is called for a unit not only when the triggered unit
of it changes state but also when a job for that unit finishes. This, firstly
allows us to make the code a bit cleaner and more generic, but more
importantly, allows us to notice correctly when a mount job fails, and
propagate that back to autofs client processes.
Fixes: #2181
|
|
job for the service queued
|
|
|
|
Also, fix indentation.
|
|
And let's make it more accurate: if we have acquire the list of unit drop-ins,
then let's do a full comparison against the old list we already have, and if
things differ in any way, we know we have to reload.
This makes sure we detect changes to drop-in directories in more cases.
|
|
This fixes fall-out from 6d10d308c6cd16528ef58fa4f5822aef936862d3.
Until that commit, do determine whether a daemon reload was required we compare
the mtime of the main unit file we loaded with the mtime of it on disk for
equality, but for drop-ins we only stored the newest mtime of all of them and
then did a "newer-than" comparison. This was brokeni with the above commit,
when all checks where changed to be for equality.
With this change all checks are now done as "newer-than", fixing the drop-in
mtime case. Strictly speaking this will not detect a number of changes that the
code before above commit detected, but given that the mtime is unlikely to go
backwards, and this is just intended to be a helpful hint anyway, this looks OK
in order to keep things simple.
Fixes: #3123
|
|
Let's move the enforcement of the per-unit start limit from unit.c into the
type-specific files again. For unit types that know a concept of "result" codes
this allows us to hook up the start limit condition to it with an explicit
result code. Also, this makes sure that the state checks in clal like
service_start() may be done before the start limit is checked, as the start
limit really should be checked last, right before everything has been verified
to be in order.
The generic start limit logic is left in unit.c, but the invocation of it is
moved into the per-type files, in the various xyz_start() functions, so that
they may place the check at the right location.
Note that this change drops the enforcement entirely from device, slice, target
and scope units, since these unit types generally may not fail activation, or
may only be activated a single time. This is also documented now.
Note that restores the "start-limit-hit" result code that existed before
6bf0f408e4833152197fb38fb10a9989c89f3a59 already in the service code. However,
it's not introduced for all units that have a result code concept.
Fixes #3166.
|
|
This way it can be used in install.c in subsequent commit.
|
|
This way it can be used in install.c in subsequent commit.
|
|
Refuse aliases to non-aliasable units in more places
Fixes #2730.
|
|
unit_has_mask_realized() determines whether the specified unit has its cgroups
set up properly given the desired target_mask; however, on the unified
hierarchy, controllers need to be enabled explicitly for children and the mask
of enabled controllers can deviate from target_mask. Only considering
target_mask in unit_has_mask_realized() can lead to false positives and
skipping enabling the requested controllers.
This patch adds unit->cgroup_enabled_mask to track which controllers are
enabled and updates unit_has_mask_realized() to also consider enable_mask.
Signed-off-by: Tejun Heo <htejun@fb.com>
|
|
If the user defines a symlink alias for a unit whose type does not support
aliasing, detect this early and print a nice warning.
Fixe: #2730
|
|
The concept of merging units exists so that we can create Unit objects for a
number of names early, and then load them only later, possibly merging units
which then turn out to be symlinked to other names. This of course only makes
sense for unit types where multiple names per unit are supported. For all
others, let's refuse the merge operation early.
|
|
We always call one after the other anyway, and this way service_set_socket_fd()
and service_close_socket_fd() nicely match each other as one undoes the effect
of the other.
|
|
Let's make sure when we drop a reference to a unit, that we run the GC queue on
it again.
This (together with the previous commit) should deal with the GC issues pointed
out in:
https://github.com/systemd/systemd/pull/2993#issuecomment-215331189
|
|
There's no need to set the no_gc bit for service units that socket units
prepare, as we always keep a proper reference (as maintained by unit_ref_set())
on them, and such references are honoured by the GC logic anyway. Moreover,
explicitly setting the no_gc bit is problematic if the socket gets GC'ed for a
reason, as the service might then leak with the bit set.
|
