| 2997173235e88c5e4cb13d2844f74afc7b25d6e2 |
|
05-Feb-2016 |
Shuqian Zhao <shuqianz@google.com> |
Add a feature to get the last UpdateAttempt ErrorCode from update_engine
For autotest, update_engine test failures are always hard to debug,
since the error message is not clear. Add a new flag 'last_attempt_error'
to get the last UpdateAttempt ErrorCode from update_engine.
Bug:25598547
Test:emerge-peppy update_engine
emerge-peppy update_engine_client
cros flash a test image to DUT.
(on the DUT):update_engine_client --last_attempt_error
Compare the results with the update_engine logs, matched.
Change-Id: Id12681097ed30b0826cad68809f17f934a07e5b2
/system/update_engine/payload_state.h
|
| 39910dcd1d68987ccee7c3031dc269233a8490bb |
|
10-Nov-2015 |
Alex Deymo <deymo@google.com> |
Split payload application code into a subdirectory.
This patch splits from the main libupdate_engine code the part that
is strictly used to download and apply a payload into a new static
library, moving the code to subdirectories. The new library is divided
in two subdirectories: common/ and payload_consumer/, and should not
depend on other update_engine files outside those two subdirectories.
The main difference between those two is that the common/ tools are more
generic and not tied to the payload consumer process, but otherwise they
are both compiled together.
There are still dependencies from the new libpayload_consumer library
into the main directory files and DBus generated files. Those will be
addressed in follow up CLs.
Bug: 25197634
Test: FEATURES=test emerge-link update_engine; `mm` on Brillo.
Change-Id: Id8d0204ea573627e6e26ca9ea17b9592ca95bc23
/system/update_engine/payload_state.h
|
| af9a863cd7aa4aa4ef12227908a10bc07440a0f0 |
|
24-Sep-2015 |
Alex Deymo <deymo@google.com> |
Remove obsolete Installer.* metrics.
These metrics were replaced by the UpdateEngine.* metrics.
Bug: chromium:355745
Test: FEATURES=test emerge-link update_engine
Change-Id: I80d0db44681c428898500dd9099918e98ae40b28
/system/update_engine/payload_state.h
|
| aea4c1cea20dda7ae7e85fc8924a2d784f70d806 |
|
20-Aug-2015 |
Alex Deymo <deymo@google.com> |
Re-license update_engine to Apache2
This patch automatically replaced the license on all text files from
Chromium OS (BSD style) to AOSP (Apache2), keeping the original year as
a reference.
The license header was added to .gyp and .gypi files, the NOTICE was
replaced with a copy of the Apache2 license and MODULE_LICENSE_* file
was updated.
BUG=b/23084294
TEST=grep 'Chromium OS Authors' doesn't find anything.
Change-Id: Ie5083750755f5180a8a785b24fe67dbf9195cd10
/system/update_engine/payload_state.h
|
| 610277efc6f7e5239158dfa4bb3b1021804326e0 |
|
12-Nov-2014 |
Alex Deymo <deymo@chromium.org> |
update_engine: Add override when possible.
Google Style Guide requires to include the "override" keyword
when overriding a method on a derived class, so the compiler will
catch errors if the method is not overriding a member of the base
class.
This patch introduces the "override" keyword when possible.
BUG=None
TEST=FEATURES=test emerge-link update_engine
Change-Id: Ie83d115c5730f3b35b3d95859a54bc1a48e0be7b
Reviewed-on: https://chromium-review.googlesource.com/228928
Tested-by: Alex Deymo <deymo@chromium.org>
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| 74b5f55e6ff608857664fe151f54d78bfe0b75bb |
|
07-Oct-2014 |
Gilad Arnold <garnold@chromium.org> |
update_engine: Relocate inference and storage of P2P related properties.
This change moves the inference of P2P related properties from
OmahaRequestAction to OmahaResponseHandlerAction, and their storage from
OmahaRequestParams to PayloadState. This is needed in order for the
UpdateCanStart policy to be able to decide P2P properties, which only
happens after the Omaha response is received and processed, and prior to
applying the update. Further, P2P properties do not affect the Omaha
request, and so there's no reason for them to reside in
OmahaRequestParams nor decided as early as OmahaRequestAction.
Additional cleanup includes swapping expected/actual arguments to EXPECT
macros where appropriate, and removing redundant .Times(1) expectation
qualifiers.
BUG=chromium:384087
TEST=Unit tests.
Change-Id: I6d5b4b44745d5dab7e350bdf019dbf804bf196a1
Reviewed-on: https://chromium-review.googlesource.com/223618
Reviewed-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
/system/update_engine/payload_state.h
|
| 519cfc71f0a2f7f01adb51acd19b5cd23d2c93a2 |
|
02-Oct-2014 |
Gilad Arnold <garnold@chromium.org> |
update_engine: Move scattering wait period store/load to PayloadState.
This is needed for decoupling the inference/use of this value from its
storing/loading, as we shift the former into the Update Manager.
BUG=chromium:384087
TEST=Unit tests.
Change-Id: I4b278dc817b6f148d5638122f934e9d1e280bfae
Reviewed-on: https://chromium-review.googlesource.com/221250
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
Reviewed-by: Alex Deymo <deymo@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
/system/update_engine/payload_state.h
|
| 6e15aace93982e885181100b2380db7fe7e1278f |
|
02-Oct-2014 |
Gilad Arnold <garnold@chromium.org> |
update_engine: Relocate data member declaration.
Per style, data members come after all method declarations.
BUG=None
TEST=None
Change-Id: Ica1d6a3c5e150d32d7386c124b8824a962295d63
Reviewed-on: https://chromium-review.googlesource.com/221199
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
/system/update_engine/payload_state.h
|
| cf175a098081f3f0e9ca52d997a7ce1585c14c2d |
|
11-Jul-2014 |
Gilad Arnold <garnold@chromium.org> |
Fix cpplint errors.
The only non-obvious change here is the switch from dynamic_cast to
static_cast in three cases of down-casting in UpdateAttempter.
dynamic_cast is banned by style, nor does it add any safety in this
particular case (subsequent code dereferences the result right away
without checking whether it's null).
BUG=None
TEST=None
Change-Id: I9d49b46362feaf9c6fa13b2715ebe9fe50308a9a
Reviewed-on: https://chromium-review.googlesource.com/207470
Tested-by: Gilad Arnold <garnold@chromium.org>
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
/system/update_engine/payload_state.h
|
| d2779df63aaad8b65fc5d4badee7dbc9bed7f2b6 |
|
16-Jun-2014 |
Alex Vakulenko <avakulenko@chromium.org> |
update_engine: fixed warnings from cpplint
Fixed all the cpplint warnings in update engine.
BUG=None
TEST=Unit tests still pass.
Change-Id: I285ae858eec8abe0b26ff203b99a42a200ceb71c
Reviewed-on: https://chromium-review.googlesource.com/204027
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Tested-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Alex Vakulenko <avakulenko@chromium.org>
/system/update_engine/payload_state.h
|
| 4e1d1495c034c8a81c989622481b6cdf94a50c3c |
|
25-Apr-2014 |
David Zeuthen <zeuthen@chromium.org> |
metrics: Emit kAbnormalTermination if last update attempt failed.
This allows us to track how many update attempts terminates
abnormally, for example if the device is rebooted in the middle of an
update or if update_engine crashes. We use a marker state variable
(aka a "prefs" file) to keep track of this.
Note that we don't currently report any of the other metrics in the
UpdateEngine.Attempt.* namespace. If necessary, this can be changed in
the future - I left a TODO item in the code with more details.
BUG=chromium:357676
TEST=New unit tests + Unit tests pass + Manual tests.
Change-Id: I83fe284c7c46917c0c55b92314c58098e2fd1789
Reviewed-on: https://chromium-review.googlesource.com/197175
Reviewed-by: Alex Deymo <deymo@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| afed4a1a355c8416d496901234932e99864d6215 |
|
10-Apr-2014 |
David Zeuthen <zeuthen@chromium.org> |
Don't report metrics when rolling back.
While reviewing histograms for the new metrics, some of the values for
the UpdateEngine.Attempt.* metrics had bogus values. For example, the
DownloadSource and ConnectionType metrics had values in their overflow
buckets. This is weird as the code carefully tries to ensure that
values outside the respective enum classes are never used.
Here's how this can happen: If we're rolling back, the UpdateAttempter
class calls PayloadState::Rollback() instead of PayloadState::UpdateResumed()
or PayloadState::UpdateRestarted().
Crucially, PayloadState::Rollback() never calls the AttemptStarted()
method so the attempt_*_ member variables are left uninitialized.
Then later on UpdateAttempter::ProcessingDone() calls
PayloadState::UpdateSucceeded() or PayloadState::UpdateFailed() which
ends up calling PayloadState::CollectAndReportAttemptMetrics() and we
report the uninitialized attempt_*_ members.
This CL fixes this problem by making PayloadState::Rollback() call
AttemptStarted() and then keep track of whether it's a rollback or
not. In the affirmative we don't report metrics. There's also a unit
test to verify this.
Additionally, this CL also fixes the oversight that the attempt_*_
members were not initialized in the constructor, per policy.
It would probably be better if rollback was implemented in another way
(so it didn't trigger codepaths like this) but that's not how it was
done. We should probably also think about reporting metrics specific
to rollback.
BUG=chromium:355745
TEST=New unit test + Unit tests pass.
Change-Id: Id2e606f02797714520290c4cbe34a056ccdae053
Reviewed-on: https://chromium-review.googlesource.com/193950
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| b281f0768f4d95afcf7378229a1a9f9ec9bd69ab |
|
02-Apr-2014 |
David Zeuthen <zeuthen@chromium.org> |
Add UpdateEngine.Attempt.ConnectionType metric.
This adds a new metric so we can track the how the device is connected
to the Internet when an attempt starts.
BUG=chromium:358339
TEST=New unit test + unit tests pass.
Change-Id: Ic5c2f50e2396e6baa288aca70906f7112ef7bca9
Reviewed-on: https://chromium-review.googlesource.com/192864
Reviewed-by: Chris Sosa <sosa@chromium.org>
Reviewed-by: Alex Deymo <deymo@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| 33bae491eded4ef4f1eb4f4ef0f01ef0e5463f3a |
|
26-Feb-2014 |
David Zeuthen <zeuthen@chromium.org> |
Add new metrics.
The current metrics (Installer.* namespace) have several shortcomings,
for example it's not immediately clear when and how frequent each
metric is reported. This CL introduces new metrics that addresses this
and other problems. The new metrics are all in the UpdateEngine.*
namespace and fall into five categories
UpdateEngine.Daily.* Reported daily.
UpdateEngine.Check.* On every check.
UpdateEngine.Attempt.* On every attempt.
UpdateEngine.SuccessfulUpdate.* With every successful update.
UpdateEngine.* Miscellaneous
Most of the new metrics mimic existing metrics and also leverage the
existing code, book-keeping and unit tests. The plan is to remove the
Installer.* metrics once we're happy with the new ones.
I've also tested this manually by performing updates and verifying
that chrome://histograms looks correct.
BUG=chromium:355745
TEST=New unit tests + unit tests pass + manual testing.
Change-Id: I7a3f68d75910384b116c7e4664776e25d3997584
Reviewed-on: https://chromium-review.googlesource.com/191314
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| 75039d7397f03dff77bdf4e26398049ff88edc4c |
|
25-Mar-2014 |
Alex Vakulenko <avakulenko@chromium.org> |
update_engine: update to libbase 242728
Made update_engine link with libchrome-242728 and fixed compile
issues due to changes in namespaces and include file layout.
Also removed some of suppressed compiler warnings and fixed the
issues masked by them (e.g. mismatched printf-like specifiers).
Added -Wextra compiler option to enable additional useful warnings
to ensure more strict checking... Had to disable "unused-parameter"
though since we have a lot of functions (mainly in fakes) that do
not use all of their parameters.
BUG=chromium:351593
TEST=Unit tests passed.
CQ-DEPEND=CL:191721
Change-Id: I1aa63a48d5f1f4ea75ba6b00aec7aa5f3bad15c4
Reviewed-on: https://chromium-review.googlesource.com/191510
Reviewed-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Alex Vakulenko <avakulenko@chromium.org>
Tested-by: Alex Vakulenko <avakulenko@chromium.org>
/system/update_engine/payload_state.h
|
| 759c275760b51defcfe5545abb887ad2616335f4 |
|
18-Mar-2014 |
Alex Deymo <deymo@chromium.org> |
Fix header guards to comply with Google Coding Style.
The Google Style Guide says that every header file should have a
define guard and the format of the symbol name should be
<PROJECT>_<PATH>_<FILE>_H_
This patch does all the minor fixes to comply with this and includes
a header guard for the bzip.h file, which didn't have it.
Also, the Copyright notice is adjusted to the Chromium OS code,
replacing "Chromium Authors" by "Chromium OS Authors".
BUG=None
TEST=build passes.
Change-Id: I6575cc307c464d60a5cb2b132cf1e46acb6500b5
Reviewed-on: https://chromium-review.googlesource.com/190445
Tested-by: Alex Deymo <deymo@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| b3dcdb309615b8ec280584319f37b187924b6694 |
|
05-Sep-2013 |
Chris Sosa <sosa@chromium.org> |
Load the and store rollback version into the correct variable.
Somehow missed this and added unittest to help never have this issue again.
In this CL I also got rid of the powerwash_prefs_ logic in GetPersistedValue
because it was just useless noise in my prev CL that isn't used anymore.
BUG=chromium:285381
TEST=unittests that actually catch error.
Change-Id: I4ed680c5b5c2e37549d58aa0a7c0d35a62623fe7
Reviewed-on: https://chromium-review.googlesource.com/168073
Tested-by: Chris Sosa <sosa@chromium.org>
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: Chris Sosa <sosa@chromium.org>
/system/update_engine/payload_state.h
|
| bb8bdc7dd7e54afe181045cc17d5a0dbde99739b |
|
03-Sep-2013 |
David Zeuthen <zeuthen@chromium.org> |
p2p: Add HttpPeer to DownloadSource enumeration
This way the Installer.DownloadSourcesUsed metric conveys how often
p2p is used and new metrics Installer.SuccessfulMBsDownloadedFromHttpPeer
and Installer.TotalMBsDownloadedFromHttpPeer gives additional detail.
BUG=chromium:284714
TEST=modify unit tests to cover this case + unit tests pass
Change-Id: Ia4dff090091a282e1a184c2b18f8290749f5fdeb
Reviewed-on: https://chromium-review.googlesource.com/167913
Reviewed-by: Alex Deymo <deymo@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| dcba8098a905018e6fe2e8e4e3b121a255f0d99e |
|
06-Aug-2013 |
David Zeuthen <zeuthen@chromium.org> |
p2p: Add accounting for p2p usage to PayloadStateInterface
This machinery is needed in order to stop using p2p if the device has
failed to update itself using p2p in a timely manner.
BUG=chromium:260426
TEST=New unit tests + unit tests pass
Change-Id: I9f33309368f8cd5399b9d67c9072a1f78385abc0
Reviewed-on: https://chromium-review.googlesource.com/64827
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| b33b0f0a0d8648542a0633cc708de149d52829b6 |
|
09-Aug-2013 |
Alex Deymo <deymo@chromium.org> |
Add Installer.UpdatesAbandonedEventCount metric.
This patch adds a new metric similar to
Installer.UpdatesAbandonedCount that reports the number of abandoned
payloads every time a payload is abandoned. A payload is considered
"abandoned" when it was already attempted but not finished for any
reason, and a newer payload is avaibled. In this situation,
update_engine drops the previous payload and starts again with the
new one.
Every time a payload is abandoned, the new metric
Installer.UpdatesAbandonedEventCount is sent with the number of
consecutive different abandoned payloads since the last successful
update. A high number of entries greater than 1 in this metric will
imply that we are releasing new payloads too often, or that the users
are taking too long to download an update.
Finally, when a payload is successfully applied, the total number of
abandoned payloads since the previous successful update is sent as
the the metric Installer.UpdatesAbandonedCount.
BUG=chromium:261370
TEST=unittests
Change-Id: I47054b541c4700a86183461a5cf92830883ea4f3
Reviewed-on: https://gerrit.chromium.org/gerrit/65328
Commit-Queue: Alex Deymo <deymo@chromium.org>
Reviewed-by: Alex Deymo <deymo@chromium.org>
Tested-by: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| 4243291a6655972fe89d4747a0089834c9e47c20 |
|
13-Jul-2013 |
Alex Deymo <deymo@chromium.org> |
Send an UMA metric when failed to boot into the new partition.
When a payload is successfully applied, the /other/ partition
is marked as valid and a reboot is needed, the reboot into this
new partition can fail due to several reasons. If than happens,
the firmware can rollback to the previous partition.
When this happens, this fix sends a new UMA metric with the
attempt number of this failing payload.
In order to test this functionality we need to fake the
utils::BootDevice() to emulate a reboot into the same or
a different partition. To achieve this, this function is
moved to a new "HardwareInterface" that can be faked
using the FakeHardware class that can hold similar hardware
related functions. Implementations and unittest were
refactored as needed.
BUG=chromium:243572
TEST=unittests
Change-Id: I1a4242df0bd61e2718ab881ead603b1d3705b877
Reviewed-on: https://gerrit.chromium.org/gerrit/61815
Commit-Queue: Alex Deymo <deymo@chromium.org>
Reviewed-by: Alex Deymo <deymo@chromium.org>
Tested-by: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| 820cc706c5ef60a3be96f5487f616bfed2b1d207 |
|
29-Jun-2013 |
Alex Deymo <deymo@chromium.org> |
Add Installer.AttemptsCount.Total metric.
This patch adds a new Installer.AttemptsCount.Total metric that
reports the number of payload attempts until the payload is
successfully applied. To achieve this a new variable persisted in
the preferences directory is added. This variable tracks the number
of attempts for the current payload, beign it a full or a delta
payload.
Unfortunatelly, there is a "PayloadAttemptNumber" variable that
tracks the number of attempts for the current payload, but only
if it is a Full payload since this value is used for the exponential
backoff algorithm used only for Full payloads. To overcome this,
this variable is renamed to "FullPayloadAttemptNumber" and the
new variable, tracking both delta and full payloads, is named
"PayloadAttemptNumber". The setter, getter and unittests are
updated as needed, and a new test is introduced to verify the
difference between these two variables when the involved payload
is a Delta payload.
BUG=chromium:226769
TEST=update_engine_unittests
Change-Id: I33d346d27ee8355f1152ebb829621571fe3a4244
Reviewed-on: https://gerrit.chromium.org/gerrit/60951
Reviewed-by: Alex Deymo <deymo@chromium.org>
Tested-by: Alex Deymo <deymo@chromium.org>
Commit-Queue: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| 1c656c48886cf5054b526ed7e95b05903d7644db |
|
28-Jun-2013 |
Alex Deymo <deymo@chromium.org> |
Add Installer.PayloadFormat metric
Adds a new PayloadFormat metric with three different values: Delta, Full
and ForcedFull. The metric is send whenever an update is applied successfully
indicating the type of the used payload. The difference between Full and
ForcedFull is that in the ForcedFull, the request sent to omaha included a
directive saying that a delta payload wasn't accepted, where a Full payload
is one where a delta payload was accepted but a Full payload was provided.
BUG=chromium:225977
TEST=cros_workon_make update_engine --test # Three tests added, one for each condition.
Change-Id: If44cd96df325d320ed327114cc5e5de3d34a5c62
Reviewed-on: https://gerrit.chromium.org/gerrit/60318
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: Alex Deymo <deymo@chromium.org>
Tested-by: Alex Deymo <deymo@chromium.org>
/system/update_engine/payload_state.h
|
| aa18e1641879f2df5993c2b02eb6f50d4443a649 |
|
20-Jun-2013 |
Chris Sosa <sosa@chromium.org> |
Blacklist versions as part of Rollback along with unittests.
This CL adds version blacklisting as part of AU Rollback. A few additional
things:
1) Since this pref must persist across rollback I have introduced a
powerwash_safe_prefs as part of system_state that will persist across
powerwashes.
2) Fixed bug where we needed to read the device policy (which is read during an
update_check before Rollback would work).
3) Some refactoring to move pref constants to constants.
4) Passing keepimg into our powerwash command so we don't wipe the old
partitions.
BUG=chromium:252589 chromium:254217
TEST=Unittests + test on device + using rollback with and without powerwash
checking preserve state.
Change-Id: I991fad944594944425fd9941e10b30a919f2b83b
Reviewed-on: https://gerrit.chromium.org/gerrit/59518
Reviewed-by: Chris Sosa <sosa@chromium.org>
Tested-by: Chris Sosa <sosa@chromium.org>
Commit-Queue: Chris Sosa <sosa@chromium.org>
/system/update_engine/payload_state.h
|
| e4c58bf01b6ccc915c6b261007dbb75241ed69b8 |
|
19-Jun-2013 |
David Zeuthen <zeuthen@chromium.org> |
Add Installer.TimeToRebootMinutes metric
This patch introduces a new metric for tracking the duration between
when an update has successfully completed (and the user is presented
with the "reboot arrow" in the panel) and when the system has booted
into the new update.
BUG=chromium:248800
TEST=New unit test + Unit tests pass + Manual tested
Change-Id: Ia22cedc3b70f1d9c2598bed9469b34a257546a64
Reviewed-on: https://gerrit.chromium.org/gerrit/59132
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Reviewed-by: Chris Sosa <sosa@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| a573d6fd046dc203b96bff6df8e57b35319906fd |
|
15-Jun-2013 |
David Zeuthen <zeuthen@chromium.org> |
Add Installer.UpdatesAbandonedCount metric
This patch adds a new metric Installer.UpdatesAbandonedCount to track
the number of update attempts that didn't complete because a newer
update was detected during the download. This is implemented by
counting the number of different responses seen since the last
successful update.
Updates are typically only abandoned if a device is suspended or
powered off while an update is happening. This can happen either
because the device was not turned on for a very long time or because
it had little or no connectivity to Omaha and/or the servers serving
the payload.
This metric will help show how many users run into this problem.
BUG=chromium:248800
TEST=New units tests + Unit tests pass + Manually tested
Change-Id: I524a380a931c2fb30916d033b7e5b0c700f57103
Reviewed-on: https://gerrit.chromium.org/gerrit/59098
Reviewed-by: Chris Sosa <sosa@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| 53173b964f6f43c89dbaba41875fcadd7077b2ea |
|
18-May-2013 |
Jay Srinivasan <jaysri@chromium.org> |
Download via HTTP only if enterprise policy allows.
In order to rollout HTTP-downloads for AU to stable channel, we want to
be a bit more conservative to preseve the defense in depth we have now
with HTTPS. So, we're introduced a new enterprise policy which should be
explicitly enabled in order for the payloads to be downloaded via HTTP.
This CL adds the support for honoring such a policy in update engine.
BUG=chromium:235562
TEST=New unit tests added, existing ones updated and they all pass.
TEST=Tested on ZGB with and without policy and it works as expected.
Change-Id: I356efbe237b10031161a57c70cb851c521915a76
Reviewed-on: https://gerrit.chromium.org/gerrit/55805
Reviewed-by: Chris Sosa <sosa@chromium.org>
Tested-by: Jay Srinivasan <jaysri@chromium.org>
Commit-Queue: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|
| a99981fda75fe0b17e96c700e3ddc93eca1cebe5 |
|
29-Apr-2013 |
David Zeuthen <zeuthen@chromium.org> |
Rename ActionExitCode to ErrorCode
Nowadays ActionExitCode is used throughout the codebase so use a more
generic name to reflect this.
BUG=chromium:216507
TEST=unit tests pass
Change-Id: I23d1d7e2676443251dbc42ed137fd018aadfa8a3
Reviewed-on: https://gerrit.chromium.org/gerrit/49512
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| 674c318a84e4344fce0fdaee039784ed7900d188 |
|
18-Apr-2013 |
David Zeuthen <zeuthen@chromium.org> |
Use better histogram parameters for duration metrics
In the current code this is capped at 20 minutes which is not helpful
as we're interested in data for devices that take much longer to
update. Instead, use one year for UpdateDuration and one month for
UpdateDurationUptime. Also move the code that sends the metric into
PayloadState to better conform with newly added code.
Also rename the metrics to they have the suffix Minutes. E.g. the new
metric names are
Installer.UpdateDurationMinutes
Installer.UpdateDurationUptimeMinutes
Also add unit test for utils::FormatTimeDelta().
BUG=None
TEST=Unit tests, look at chrome://histogram
Change-Id: I8581ec111f9833bdd72f10888c9626f40779f8b9
Reviewed-on: https://gerrit.chromium.org/gerrit/48558
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| be45bef9e283188b00e7def8967f81843669a7f1 |
|
10-Apr-2013 |
Chris Sosa <sosa@chromium.org> |
Add update reboot metric to the update engine.
This change add the Installer.UpdateNumReboots metric.
This records the number of reboots that occurred while an update was being
attempted. It uses a marker file stored in tmp to discover whether or not
it's already recorded the reboot.
BUG=chromium:226766
TEST=Unittests | ran an update on a test machine and rebooted/resumed and
checked about:histograms to confirm numbers. Also restart update-engine to
verify it didn't double count that.
Change-Id: I5d2af9d5b62a9d974c7c6243a89cb3359051b650
Reviewed-on: https://gerrit.chromium.org/gerrit/47710
Tested-by: Chris Sosa <sosa@chromium.org>
Commit-Queue: Chris Sosa <sosa@chromium.org>
Reviewed-by: Chris Sosa <sosa@chromium.org>
/system/update_engine/payload_state.h
|
| cc6f99600b3354cfe3a0d212241e1ee1dc3421b6 |
|
18-Apr-2013 |
David Zeuthen <zeuthen@chromium.org> |
Add metric to report number of URL switches per update attempt
BUG=chromium:226765
TEST=Existing unit tests modified to test feature, Unit tests pass
Change-Id: I5ee0816cac3fa4fa641949ed4872aef8074e3fcc
Reviewed-on: https://gerrit.chromium.org/gerrit/48537
Commit-Queue: David Zeuthen <zeuthen@chromium.org>
Reviewed-by: David Zeuthen <zeuthen@chromium.org>
Tested-by: David Zeuthen <zeuthen@chromium.org>
/system/update_engine/payload_state.h
|
| 19409b74019d787100b768306e75ab3e5882898d |
|
13-Apr-2013 |
Jay Srinivasan <jaysri@chromium.org> |
Add metrics to report the number of bytes downloaded per protocol.
This CL adds these two basic metrics:
1. The number of megabytes downloaded using each protocol (or source
in general, as we want to consider HTTP downloads from server and possibly
a peer in future as two different sources) for each successful update. The
totals across all protocols will equal the payload size exactly.
2. The total number of megabytes downloaded since the last update attempt.
If there are no errors, this should be same as above. Otherwise, this will
be larger and will help us to compute the update efficiency.
BUG=chromium:225953
TEST=Unit tests pass, new Unit tests added.
TEST=chrome://histograms shows metrics correctly.
Change-Id: Ic02e218f46568427df99a8a9df2011860aee84f3
Reviewed-on: https://gerrit.chromium.org/gerrit/48069
Reviewed-by: Chris Sosa <sosa@chromium.org>
Commit-Queue: Jay Srinivasan <jaysri@chromium.org>
Tested-by: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|
| 9a017f2c5ae41c04a7c7c15b5dbd08faadae7606 |
|
12-Apr-2013 |
David Zeuthen <zeuthen@chromium.org> |
Measure and send update duration (and corresponding uptime) to UMA
This patch introduces two new metrics, Installer.UpdateDuration and
Installer.UpdateDurationUptime. The former is the timespan from when
the update was first discovered until it has been downloaded and
applied (including the time the device is suspended or powered
off). The latter is similar, but without taking into account time
spent in suspend or powered off.
For example, if the device is suspended (or powered off) for N seconds
while updating, the Installer.UpdateDuration metric will be N seconds
bigger than Installer.UpdateDurationUptime metric:
Histogram: Installer.UpdateDuration recorded 1 samples, average = 313.0
Histogram: Installer.UpdateDurationUptime recorded 1 samples, average = 251.0
Also remove the existing Installer.UpdateTime metric as this didn't
take process restarts into account and is now superseeded by the
Installer.UpdateDuration metric.
This is done by using the CLOCK_MONOTONIC_RAW clock (available in
Linux 2.6.28 and later) since this clock indeed does not advance when
the system is sleeping.
We use the PayloadState class to persist recorded data across
update_engine process restart (including device reboots).
Since clock_gettime(2) and CLOCK_MONOTONIC_RAW requires linking to the
librt library do this and also request the system header files to
expose the required symbols and defines, i.e. define _POSIX_C_SOURCE
>= 199309L.
Also remove _POSIX_C_SOURCE mangling from update_attempter.cc since
it's actually not needed there and generally it's better to make the
environment the same across all translation units (by putting whatever
is needed in e.g. CCFLAGS).
BUG=chromium:226763
TEST=unit tests, force update, examine chrome://histograms
Change-Id: I883668564b5fa78ff3e19156bd77496ff929ca58
Signed-off-by: David Zeuthen <zeuthen@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/47928
Reviewed-by: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|
| 082628869ed3ee3e173c08354d3fc40cdb7df2c0 |
|
29-Dec-2012 |
Jay Srinivasan <jaysri@chromium.org> |
Implement exponential backoff for throttling repeated AU downloads.
Today we retry the same payload over and over again every hour. Ideally,
we shouldn't require ever to re-download the same payload again. But
from experience we find that post-install or firmware updates may succeed
on a second attempt. So until we have code that can do such selective
retries of those steps, we currently re-download and re-apply the whole
payload. So instead of retrying over and over again, we backoff the
successive payload download attempts at 1 day, 2 days, 4 days, etc. with
an upper limit of 16 days.
Another subtle reason for which we depend on the payload retry mechanism
today is if we've failed downloading the payload via all the URLs that are
specified in the rule, we don't want to keep re-attempting the download.
This case is different from the case discussed above, because in this case
we haven't even downloaded a payload once completely. In this case also,
there's a need for throttling the amount of bytes we end up downloading
repeatedly for a particular operation that may fail. This is done by
treating the exhaustion of all URLs as equivalent to having downloaded
a full payload and subjecting it to the same backoff behavior.
We waive backoffs for dev/test images so as not to cause any delay in
our testing or development.
BUG=chromium-os:36806
TEST=Added new unit tests. Tested all scenarios on my ZGB.
Change-Id: I6bd0d3f296a3c0da0a8026fb71b24785d825e39c
Reviewed-on: https://gerrit.chromium.org/gerrit/40220
Commit-Queue: Jay Srinivasan <jaysri@chromium.org>
Reviewed-by: Jay Srinivasan <jaysri@chromium.org>
Tested-by: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|
| 2b5a0f065187fd19179e3809148dbfc376ada7a0 |
|
20-Dec-2012 |
Jay Srinivasan <jaysri@chromium.org> |
Classify errors and advance URL index according to the error code.
In CL https://gerrit.chromium.org/gerrit/39638, we always incremented
the URL index irrespective of the error code. That would cause the first
URL to be given up too quickly in favor of the second one even for
transient errors such as when user closes a lid and reopens after some
time.
The right behavior in this case is to just count those failures towards
the URL and only after repeated failures with no progress should we
advance the URL index.
This CL implements this logic and completes the multiple URL-related
work items outlined in the design doc.
BUG=chromium-os:37206
TEST=Tested all uses cases on my ZGB. Added and updated unit tests.
Change-Id: Ida0cfbfeb9bfab732144049d1b27e3b8958bc252
Reviewed-on: https://gerrit.chromium.org/gerrit/39885
Commit-Queue: Jay Srinivasan <jaysri@chromium.org>
Reviewed-by: Jay Srinivasan <jaysri@chromium.org>
Tested-by: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|
| 6f6ea00aa8c4cf54b6842be32ca1226854c24f78 |
|
14-Dec-2012 |
Jay Srinivasan <jaysri@chromium.org> |
Support for processing multiple URLs in update_engine.
Main changes:
1. Added a new PayloadState class which encapsulates all the persisted
state we use for multiple URLs, back-off (TBD), etc.
2. Added support for handling multiple URLs stored in the OmahaResponse in
OmahaRequestAction and OmahaResponseHandlerAction code.
3. Added support for picking the right URL in OmahaResponseHandlerAction
and putting it in the install_plan. This way, the rest of the code that
uses the install_plan is oblivious to the presence of multiple URLs :-)
4. Added support for advancing to next URL when an update fails. The full
error classification is a new work item (chromium-os:37206). Right now,
it's a basic round-robin on every error.
5. Updated the conditions for determining when hash checks are mandatory.
Previously since there was only one URL, if it was HTTPS, the checks were
waived. Now, even if there's one HTTP URL, we make hash checks mandatory
even if other HTTPS URLs are present.
6. Added new unit tests for PayloadState and the new logic added to other
places.
Noisy changes:
1. Instead of passing PrefsInterface to OmahaRequestAction and
OmahaResponseHandlerAction, we're now passing SystemState which will now
contain PrefsInterface and the newly added PayloadState object that these
actions need to do their work.
2. Renamed a bunch of setters/getters to set_x() and x() instead of SetX()
and GetX() methods - this was pending from Gilad's old CR. As I'm
adding new methods in the correct style, I went ahead and fixed it to
avoid the confusing styles.
3. Updated all existing unit tests to reflect these changes.
BUG=chromium-os:36807
TEST=All Single/Multiple URL scenarios work fine on my ZGB as expected.
TEST=Old and new unit tests run fine.
Change-Id: Id31f9ccb220471f3ec3a475f624dc03c16119144
Reviewed-on: https://gerrit.chromium.org/gerrit/39638
Commit-Ready: Jay Srinivasan <jaysri@chromium.org>
Reviewed-by: Jay Srinivasan <jaysri@chromium.org>
Tested-by: Jay Srinivasan <jaysri@chromium.org>
/system/update_engine/payload_state.h
|