[PATCH] cooker: limit the execution of the hashequiv client/server ping

[PATCH] cooker: limit the execution of the hashequiv client/server ping

[corentin.guillevic]

From: Corentin Guillevic <corentin.guillevic@smile.fr>

The ping method is used to test the connection to the hash equivalence server.
However, if the server is unavailable, the bitbake server has to wait several
minutes for the system to terminate.

While waiting for a response from the hash server, the bitbake server does not
provide any feedback to the bitbake client. The latter then closes the link and
fails with the error:

Timeout while waiting for a reply from the bitbake server (60s at 09:50:15.377337)

Adding a timeout to the ping step allows quicker action when a remote server is
unavailable, and prevents disconnection of the client due to timeout.
The default timeout is 10 seconds, but this can be set using the
BB_HASHSERVER_TIMEOUT variable.

Signed-off-by: Corentin Guillevic <corentin.guillevic@smile.fr>
---
 lib/bb/asyncrpc/client.py | 4 ++--
 lib/bb/cooker.py          | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/bb/asyncrpc/client.py b/lib/bb/asyncrpc/client.py
index 17b72033b..7ef76c90c 100644
--- a/lib/bb/asyncrpc/client.py
+++ b/lib/bb/asyncrpc/client.py
@@ -199,8 +199,8 @@ class AsyncClient(object):
         self.check_invoke_error(result)
         return result
 
-    async def ping(self):
-        return await self.invoke({"ping": {}})
+    async def ping(self, timeout):
+        return await asyncio.wait_for(self.invoke({"ping": {}}), timeout)
 
     async def __aenter__(self):
         return self
diff --git a/lib/bb/cooker.py b/lib/bb/cooker.py
index 4b6ba3196..0bc9b7aec 100644
--- a/lib/bb/cooker.py
+++ b/lib/bb/cooker.py
@@ -342,7 +342,7 @@ according to https://docs.yoctoproject.org/dev-manual/hashequivserver.html""".fo
                 if upstream:
                     try:
                         with hashserv.create_client(upstream) as client:
-                            client.ping()
+                            client.ping(self.data.getVar("BB_HASHSERVER_TIMEOUT") or 10)
                     except ImportError as e:
                         bb.fatal(""""Unable to use hash equivalence server at '%s' due to missing or incorrect python module:
 %s
@@ -352,7 +352,7 @@ Please install the needed module on the build host, or use an environment contai
  - or set up pip venv
 You can also remove the BB_HASHSERVE_UPSTREAM setting, but this may result in significantly longer build times as bitbake will be unable to reuse prebuilt sstate artefacts."""
                                  % (upstream, repr(e)))
-                    except ConnectionError as e:
+                    except (ConnectionError, TimeoutError) as e:
                         bb.warn("Unable to connect to hash equivalence server at '%s', please correct or remove BB_HASHSERVE_UPSTREAM:\n%s"
                                  % (upstream, repr(e)))
                         upstream = None
-- 
2.51.0

Re: [bitbake-devel] [PATCH] cooker: limit the execution of the hashequiv client/server ping

[Antonin Godard]

Hi,

On Thu Apr 23, 2026 at 12:29 PM CEST, Corentin Guillevic via lists.openembedded.org wrote:
> From: Corentin Guillevic <corentin.guillevic@smile.fr>
>
> The ping method is used to test the connection to the hash equivalence server.
> However, if the server is unavailable, the bitbake server has to wait several
> minutes for the system to terminate.
>
> While waiting for a response from the hash server, the bitbake server does not
> provide any feedback to the bitbake client. The latter then closes the link and
> fails with the error:
>
> Timeout while waiting for a reply from the bitbake server (60s at 09:50:15.377337)
>
> Adding a timeout to the ping step allows quicker action when a remote server is
> unavailable, and prevents disconnection of the client due to timeout.
> The default timeout is 10 seconds, but this can be set using the
> BB_HASHSERVER_TIMEOUT variable.

Can you please document this new variable in
doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst in the same
commit?

Thanks,
Antonin

Re: [bitbake-devel] [PATCH] cooker: limit the execution of the hashequiv client/server ping

[Mathieu Dubois-Briand]

On Thu Apr 23, 2026 at 12:29 PM CEST, Corentin Guillevic via lists.openembedded.org wrote:
> From: Corentin Guillevic <corentin.guillevic@smile.fr>
>
> The ping method is used to test the connection to the hash equivalence server.
> However, if the server is unavailable, the bitbake server has to wait several
> minutes for the system to terminate.
>
> While waiting for a response from the hash server, the bitbake server does not
> provide any feedback to the bitbake client. The latter then closes the link and
> fails with the error:
>
> Timeout while waiting for a reply from the bitbake server (60s at 09:50:15.377337)
>
> Adding a timeout to the ping step allows quicker action when a remote server is
> unavailable, and prevents disconnection of the client due to timeout.
> The default timeout is 10 seconds, but this can be set using the
> BB_HASHSERVER_TIMEOUT variable.
>
> Signed-off-by: Corentin Guillevic <corentin.guillevic@smile.fr>
> ---

Hi Corentin,

Thanks for your patch.

It looks like this is triggering some oe-selftest failures:

2026-04-24 13:28:03,744 - oe-selftest - INFO - prservice.BitbakePrTests.test_import_export_override_db (subunit.RemotedTestCase)
2026-04-24 13:28:03,746 - oe-selftest - INFO -  ... FAIL
...
  File "/srv/pokybuild/yocto-worker/oe-selftest-armhost/build/layers/openembedded-core/meta/lib/oeqa/selftest/cases/prservice.py", line 115, in test_import_export_override_db
    self.run_test_pr_export_import('m4', replace_current_db=False)
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/srv/pokybuild/yocto-worker/oe-selftest-armhost/build/layers/openembedded-core/meta/lib/oeqa/selftest/cases/prservice.py", line 89, in run_test_pr_export_import
    self.increment_package_pr(package_name)
    ~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^
  File "/srv/pokybuild/yocto-worker/oe-selftest-armhost/build/layers/openembedded-core/meta/lib/oeqa/selftest/cases/prservice.py", line 61, in increment_package_pr
    self.assertEqual(res.status, 0, msg=res.output)
    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.13/unittest/case.py", line 907, in assertEqual
    assertion_func(first, second, msg=msg)
    ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.13/unittest/case.py", line 900, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: 1 != 0 : NOTE: Reconnecting to bitbake server...
NOTE: Retrying server connection (#1)... (13:28:02.531366)
NOTE: Started PRServer with DBfile: /srv/pokybuild/yocto-worker/oe-selftest-armhost/build/build-st-1371670/cache/prserv.sqlite3, Address: 127.0.0.1:37647, PID: 2273277
ERROR: PRservice localhost:37647 not available
ERROR: Unable to start PR Server, exiting, check the bitbake-cookerdaemon.log
...
2026-04-24 13:28:06,081 - oe-selftest - INFO - prservice.BitbakePrTests.test_import_export_replace_db (subunit.RemotedTestCase)
2026-04-24 13:28:06,081 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:08,453 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_deb_arch_dep (subunit.RemotedTestCase)
2026-04-24 13:28:08,453 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:10,791 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_deb_arch_indep (subunit.RemotedTestCase)
2026-04-24 13:28:10,792 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:13,383 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_ipk_arch_dep (subunit.RemotedTestCase)
2026-04-24 13:28:13,384 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:15,719 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_ipk_arch_indep (subunit.RemotedTestCase)
2026-04-24 13:28:15,720 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:18,037 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_rpm_arch_dep (subunit.RemotedTestCase)
2026-04-24 13:28:18,038 - oe-selftest - INFO -  ... FAIL
...
2026-04-24 13:28:20,344 - oe-selftest - INFO - prservice.BitbakePrTests.test_pr_service_rpm_arch_indep (subunit.RemotedTestCase)
2026-04-24 13:28:20,345 - oe-selftest - INFO -  ... FAIL

https://autobuilder.yoctoproject.org/valkyrie/#/builders/23/builds/3834
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/3744
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/3624

Can you have a look at the issue?

Thanks,
Mathieu

-- 
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com