[BUG]: After upgrade to EOS connect version 0.3.35 - problems with HA sensors (could not convert string to float) and '.../packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'

[frog60]

Hopefully I am in the right git project!?

What happend:

upgrading to using EOS connect version 0.3.35
I am using "local_evopt" as optimizer
HA is installed as VM in a proxmos environment (no change)
After installing the new EOS connect version 0.3.35 I can see the following errors in the log (see below for details:

2026-06-23 18:58:46 ERROR [OPT-LocalEVopt] Optimization failed: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'

and for HA configured sensor data (see below for details):

...
2026-06-23 18:46:06 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:08:25: state '42.602' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A08%3A25.460464%2B00%3A00&end_date=2026-06-09T10%3A08%3A25.460464%2B00%3A00)
...

In the EOS connect version 0.2.34 before I used the docker image as optimization engine. It worked in that setup.

Is there a solution on my issues? I checked the doc, but I did not find any hints on these issues.

Best regards

Michael

...
2026-06-23 18:58:21 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:09:49: state '' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A09%3A49.381684%2B00%3A00&end_date=2026-06-09T10%3A09%3A49.381684%2B00%3A00)
2026-06-23 18:58:21 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:10:11: state '45.799' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A10%3A11.462562%2B00%3A00&end_date=2026-06-09T10%3A10%3A11.462562%2B00%3A00)
2026-06-23 18:58:21 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:10:19: state '' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A10%3A19.448762%2B00%3A00&end_date=2026-06-09T10%3A10%3A19.448762%2B00%3A00)
2026-06-23 18:58:21 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:13:05: state '46.089' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A13%3A05.228817%2B00%3A00&end_date=2026-06-09T10%3A13%3A05.228817%2B00%3A00)
2026-06-23 18:58:21 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-09 08:13:07: state '' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-09T06%3A13%3A07.317297%2B00%3A00&end_date=2026-06-09T10%3A13%3A07.317297%2B00%3A00)
...

...
2026-06-23 18:58:46 ERROR [OPT-LocalEVopt] Optimization failed: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'
Traceback (most recent call last):
File "/app/src/interfaces/optimization_backends/optimization_backend_local_evopt.py", line 141, in optimize
evopt_response = optimizer.solve()
File "/app/src/interfaces/optimization_backends/local_evopt/optimizer.py", line 716, in solve
self.problem.solve(solver)
~~~~~~~~~~~~~~~~~~^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/pulp.py", line 2407, in solve
status = solver.actualSolve(self, **kwargs)
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 406, in actualSolve
return super().actualSolve(lp, **kwargs)
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 153, in actualSolve
return self.solve_CBC(lp, **kwargs)
~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 230, in solve_CBC
cbc = subprocess.Popen(args, stdout=pipe, stderr=pipe, stdin=devnull)
File "/usr/local/lib/python3.13/subprocess.py", line 1039, in init
self._execute_child(args, executable, preexec_fn, close_fds,
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
pass_fds, cwd, env,
^^^^^^^^^^^^^^^^^^^
...<5 lines>...
gid, gids, uid, umask,
^^^^^^^^^^^^^^^^^^^^^^
start_new_session, process_group)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.13/subprocess.py", line 1991, in _execute_child
raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'
...

[ohAnd]

check #260
and https://ohand.github.io/EOS_connect/user-guide/index.html#troubleshooting
or "Note for Proxmox / VM Users:" in https://github.com/ohAnd/EOS_connect/blob/main/README.md

[treffer-priv]

Same issue here after upgrading to 0.3.35, using local_evopt as optimizer.

Environment: Home Assistant OS 18.0 (not Proxmox) on x86-host

Error:

Optimization failed: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'

Troubleshooting steps tried (all unsuccessful):

1. Uninstalled the add-on (including app data) and reinstalled from scratch
2. Manually removed the Docker image as described in the documentation:
   docker rmi ghcr.io/ohand/ha-addon-eos_connect_amd64:0.3.35
   and reinstalled — error persists

I did some digging inside the running container and I believe I found the root cause. Note: the initial hypothesis was suggested by an AI assistant, but the findings below are based on actual commands run inside the container.

What I checked:

The CBC binary is actually present and has correct permissions:

-rwxr-xr-x 1 root root 7178823 Jun 23 14:55 /opt/venv/lib/python3.13/site-packages/pulp/solverdir/cbc/linux/i64/cbc

But it cannot execute. The binary requires glibc:

$ strings .../cbc | grep "^/lib"
/lib64/ld-linux-x86-64.so.2

While the container is Alpine-based and only has musl:

$ ls /lib/ld*
/lib/ld-musl-x86_64.so.1
$ ls /lib64/
ls: /lib64/: No such file or directory

I tried both Alpine glibc compatibility layers (gcompat and libc6-compat), both fail with the same error:

Error relocating .../cbc: __fpu_control: symbol not found

__fpu_control is a glibc-internal symbol that musl does not implement — no compatibility shim can fix this.

AI Conclusion: The pulp CBC binary is compiled against glibc but the container base image is Alpine (musl). These are fundamentally incompatible. We could not simulate a working fix locally, but the evidence is clear: switching the base image to Debian/Ubuntu (which ships glibc) should resolve this.

Workaround for now: Switching the optimizer to an external EOS (evopt) instance works without errors.
So it seems the CBC solver binary is simply missing from (or incompatible with) the add-on image when running on HA OS. Happy to provide any additional logs or info if helpful.

Hope it will help finding a solution.

[ohAnd]

investigated now a lot ... for this "sorry" edge case ;-)

back to the older problem which is already described and to be sure this is already set at your side
=>

Note for Proxmox / VM Users: If the add-on crashes with a Segmentation Fault on startup, your VM might be using a generic CPU type.

Go to your VM Settings > Hardware > Processor.
Change Type from kvm64 (default) to host.
Restart the VM.

Please check if already done at your side!

---

the AI conclusion sounds reasonable at the first glance, but with my understanding (now ;-) )
-> alpine linux is the base for all provided HA addon packages (AMD64/ Aarch64)

• the host system provide a system environment that can be handled by a "modern" linux
• at proxmox the kvm64 seems the deafult after install and is an older generic emulation which has not the needed capability
• so the switch to use host environment provide the right environment (CPU instructionn set ) that this lib is useable

PS: most users are using the HA addon/ app of EOS_connect - currently at around 400 for the last release. Given that, I would expect there to be more complaints if there were bigger issues ;-) (and during test phase for develop version in avg ~ 150)

[frog60]

@ohAnd : Thx for feedback. I already did the Proxmos VM Settings > Hardware > Processor

But I am still not able to get a successful response from the optimize request.

What I can see in the logs:

• some sensors (here: sensor.total_home_power) with "corrupted data" - no idea on that.
• 2026-06-26 16:17:25 ERROR [OPT-EOS] OPTIMIZE Request failed: 422 Client Error: Unprocessable Content for url: http://192.168.178.193:8503/optimize?start_hour=16

EOS server ist running - with EOS Connect 0.2.34 the connection to the EOS server works. Nothing changed in any port configuration ...

Calling: http://192.168.178.193:8503/v1/config

I get the response - does the API of eos_server (akkudocktor) change and the endpoint "http://192.168.178.193:8503/optimize?start_hour=16" is wrong?

So I tried to roll back to EOS connect 0.2.34 using the HA backup, but I failed with:

I stuck here. Sorry.

This is the output for configuring eos_server:

...

2026-06-26 16:17:05 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.total_home_power' at 2026-06-12 19:18:42: state '600.4' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.total_home_power&start_date=2026-06-12T17%3A18%3A42.767103%2B00%3A00&end_date=2026-06-12T21%3A18%3A42.767103%2B00%3A00)
2026-06-26 16:17:05 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.total_home_power' at 2026-06-12 19:32:05: state '539.2' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.total_home_power&start_date=2026-06-12T17%3A32%3A05.976611%2B00%3A00&end_date=2026-06-12T21%3A32%3A05.976611%2B00%3A00)
2026-06-26 16:17:25 INFO [CHARGE_DEMAND] Optimizer request preparation: max_charge_power=1350 W (source=dynamic, dynamic_max=1350, config_fixed=8400, charging_curve_enabled=True)
2026-06-26 16:17:25 INFO [OPT-EOS] OPTIMIZE request optimization with: http://192.168.178.193:8503/optimize?start_hour=16 - and with timeout: 180
2026-06-26 16:17:25 INFO [OPT-EOS] OPTIMIZE response retrieved successfully in 0 min 0.01 sec for current run
2026-06-26 16:17:25 ERROR [OPT-EOS] OPTIMIZE Request failed: 422 Client Error: Unprocessable Content for url: http://192.168.178.193:8503/optimize?start_hour=16
2026-06-26 16:17:25 ERROR [OPT-EOS] OPTIMIZE Response status: 422
2026-06-26 16:17:25 WARNING [Main] optimize() returned no avg_runtime; keeping previous value: 120
2026-06-26 16:17:25 ERROR [OPTIMIZATION] No control data in optimized response
2026-06-26 16:17:25 INFO [OPTIMIZATION] Gap-fill run: start 16:20:25 (quarter-aligned run follows at 16:30:00)
2026-06-26 16:17:25 INFO [Main] Next optimization at 16:20:25 (based on average runtime of 120 seconds). Sleeping for 3 min 0 seconds
2026-06-26 16:17:26 INFO [CHARGE_DEMAND] Updating slot power (reasons: new_slot): demand=0 Wh, slot=2554s, calculated=0 W, capped=0 W
2026-06-26 16:17:26 INFO [CHARGE_DEMAND] Slot power updated to 0 W (new commitment)
2026-06-26 16:17:26 INFO [Main] Inverter mode set to MODE AVOID DISCHARGE, PV charge rate: 1350 W (-----)

[frog60]

@ohAnd configuring local_evopt

I still have the same issues:

2026-06-26 17:00:32 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.heat_pump_total_power' at 2026-06-12 19:32:07: state '' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.heat_pump_total_power&start_date=2026-06-12T17%3A32%3A07.192343%2B00%3A00&end_date=2026-06-12T21%3A32%3A07.192343%2B00%3A00)
2026-06-26 17:00:32 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.total_home_power' at 2026-06-12 19:11:48: state '506.5' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.total_home_power&start_date=2026-06-12T17%3A11%3A48.362027%2B00%3A00&end_date=2026-06-12T21%3A11%3A48.362027%2B00%3A00)
2026-06-26 17:00:32 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.total_home_power' at 2026-06-12 19:18:42: state '600.4' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.total_home_power&start_date=2026-06-12T17%3A18%3A42.767103%2B00%3A00&end_date=2026-06-12T21%3A18%3A42.767103%2B00%3A00)
2026-06-26 17:00:33 INFO [LOAD-IF] Skipping invalid sensor data for 'sensor.total_home_power' at 2026-06-12 19:32:05: state '539.2' cannot be processed (could not convert string to float: ''). This may indicate missing or corrupted data in the database. (check: http://192.168.178.114:8123/history?entity_id=sensor.total_home_power&start_date=2026-06-12T17%3A32%3A05.976611%2B00%3A00&end_date=2026-06-12T21%3A32%3A05.976611%2B00%3A00)
2026-06-26 17:00:38 INFO [CHARGE_DEMAND] Optimizer request preparation: max_charge_power=1350 W (source=dynamic, dynamic_max=1350, config_fixed=8400, charging_curve_enabled=True)
2026-06-26 17:00:38 WARNING [OPT-EVopt] initial_soc (26626.60 Wh, 98.00%) > s_max (26354.90 Wh, 97.00%); clamping to s_max
2026-06-26 17:00:38 ERROR [OPT-LocalEVopt] Optimization failed: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'
Traceback (most recent call last):
File "/app/src/interfaces/optimization_backends/optimization_backend_local_evopt.py", line 141, in optimize
evopt_response = optimizer.solve()
File "/app/src/interfaces/optimization_backends/local_evopt/optimizer.py", line 716, in solve
self.problem.solve(solver)
~~~~~~~~~~~~~~~~~~^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/pulp.py", line 2407, in solve
status = solver.actualSolve(self, **kwargs)
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 406, in actualSolve
return super().actualSolve(lp, **kwargs)
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 153, in actualSolve
return self.solve_CBC(lp, **kwargs)
~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/opt/venv/lib/python3.13/site-packages/pulp/apis/coin_api.py", line 230, in solve_CBC
cbc = subprocess.Popen(args, stdout=pipe, stderr=pipe, stdin=devnull)
File "/usr/local/lib/python3.13/subprocess.py", line 1039, in init
self._execute_child(args, executable, preexec_fn, close_fds,
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
pass_fds, cwd, env,
^^^^^^^^^^^^^^^^^^^
...<5 lines>...
gid, gids, uid, umask,
^^^^^^^^^^^^^^^^^^^^^^
start_new_session, process_group)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.13/subprocess.py", line 1991, in _execute_child
raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: '/opt/venv/lib/python3.13/site-packages/pulp/apis/../solverdir/cbc/linux/i64/cbc'
2026-06-26 17:00:38 WARNING [Main] optimize() returned no avg_runtime; keeping previous value: 120
2026-06-26 17:00:38 ERROR [OPTIMIZATION] No control data in optimized response

[ohAnd]

@ohAnd : Thx for feedback. I already did the Proxmos VM Settings > Hardware > Processor

But I am still not able to get a successful response from the optimize request.

maybe dumb question - whole server / hardware/ VM restarted after changed kvm64 -> host? (sorry, but the ideas for a solution are decreasing ;-) )

second problem (EOS optimizer with 422 answer) is a fully another one and not connected with CBC - please move to an new issue

[frog60]

@ohAnd: added addition bug #267 as requested for the http error 422.

BTW: In Proxmos environoment the complete Proxmos environment was restarted after changing processor to "host".