I run my pytorch code well on mac and even on windows system but the same code seems stuck on CentOS6.3.
I debug with ipdb, and found the code was stuck at F.conv2d function:
> /home/work/anaconda2/envs/PyTorch/lib/python2.7/site-packages/torch/nn/modules/conv.py(301)forward()
300 return F.conv2d(input, self.weight, self.bias, self.stride,
--> 301 self.padding, self.dilation, self.groups)
302
ipdb> s
The running env was created with anaconda(python 2.7/3.6), pytorch version is 0.4.0.
I tried for a long time to resolve this problem and i tried. Do you have a suggestion? Thank you so much!
I reinstall CentOS6.3, and then upgrade glibc2.14, glibc2.17 due to the pytorch0.4.0 running error info.
Now everything is ok.
By the way, the pytorch0.3.1 perform well before i upgrade the glibc(up to 2.12). So i think the lastest pytorch0.4.0 may haven’t deal very well with glibc, leave running deadlock appearance and doesn’t tell any error and warning info, just stuck at F.conv2d in torch/nn/modules/conv.py(301).
See also: https://discuss.pytorch.org/t/f-conv2d-stuck-on-my-centos/19794/3
Related
I'm using a portable install of VS Code 1.56.1 on Windows 10. Whenever open an extension from the marketplace, I get the following error:
Error loading webview: Error: Could not register service workers: TypeError: Failed to register a ServiceWorker for scope ('vscode-webview://extensioneditor/') with script ('vscode-webview://extensioneditor/service-worker.js?platform=electron&id=extensionEditor&vscode-resource-origin=https%3A%2F%2FextensionEditor.vscode-webview-test.com'): ServiceWorker cannot be started.
Same thing happens for other things using webviews, with a slightly different path (like VS Code release notes). I'm able to install plugins and view the Feature Contributions tab, but I cannot view the Details and Changelog tabs.
Any thoughts on what could be causing this? This is on a managed machine where security settings might be changing. I'm using a portable install because my installed version stopped working out of the blue last week and this was supposed to be the workaround until that was figured out.
I was getting the same error, it looks like an issue introduced on version 1.56.0.
For me the fix was running VS Code without elevated permissions.
I always run as administrator and running without it fixed the issue for me.
I had this problem, in terminal use code . --no-sandbox
Updating python3 to the most recent version may also help.
I created a solution with a few applications, and it worked perfectly. But a day later when I was about to debug the solution again, it suddenly cannot start.
I did have firewall and AV disabled, both when it worked and later when it didn't work.
It works fine when deployed to azure, but not locally.
The error thrown is:
Exception from HRESULT: 0x80071BFF
at System.Fabric.Interop.NativeRuntime.FabricGetNodeContext()
at System.Fabric.FabricRuntime.NativeFabricRuntimeFactory.GetNodeContextHelper()
at System.Fabric.Interop.Utility.WrapNativeSyncInvoke[TResult](Func`1 func, String functionTag, String functionArgs)
A related post register-servicefabricapplicationtype-on-a-secure-cluster-always-times-out
describes a similar thing.
However, I get this while debugging locally and with smallest possible application: I even just created 1 application, 1 actor, did no changes, hit F5 and I get this error.
So, I looked at service-fabric-troubleshoot-local-cluster-setup and while I also get the TypeInitializationException, the solution of:
Your path variable was not correctly set during installation. Please sign out of Windows and sign back in. This will fully refresh your path.
did not work at all.
Nothing else on that page seemed to be related.
Now I begin to feel that I've hit the end of what key words I can google, and still I have no idea what to do.
Any help would be greatly appreciated.
EDIT:
I've tried starting via cluster manager and apps can still not start.
I've removed apps, removed cluster, tried with new solutions, over and over. The only thing I haven't tried is reinstalling SDK and VS. I'm a bit reluctant to reinstall VS, so I'd hope to find some clues before resorting to that.
UPDATE1:
Now I reinstalled Service Fabric v. 5.3.311.9590, SDK and Tools.
Created a solution, added a stateless service. Hit F5. The exact same error is thrown.
I will now try to install on another machine with mostly the same configuration.
UPDATE2:
Installed the SDK on an identical VM, Win10, also with VS2015 Update 3. Created solution, added 1 actor, hit F5 and also on this macchine the exact error is thrown.
I tried Debug without debugging (as mentioned here) to attach later. But the application never starts. It is failed. This is the same on both the machines.
It all worked and from one day to the other it doesn't, and it's a problem that I cannot find anywhere on the net. What can this be? I found that the security updates from windows was made about the same date this happened...
I will uninstall and try again.
UPDATE3:
Uninstalling security update was not possible, but I could hide a couple of other updates. To no luck though.
From here I found this https://github.com/Azure/service-fabric-issues/issues/15 and realised I actually was very low on diskspace, and so I increased it (hyper-v manager) to 20 gigs free. But no, still the same problem.
I have an issue with nodeclipse: I am unable to run even the simplest demo app (simple require http). When I try to run my "app.js" as a node application, I am just getting:
<terminated, exit value: -1073741819>Node.js Process
When I try to run from the command prompt node app.js, everything goes fine. Did I forget anything important ?
Thanks in advance for your help.
Here is some additionnal informations:
Env config: Win 8 64 bits + Eclipse Kepler 64 bits (JRE 6.0.390.4)
Enide from Eclipse marketplace (based on Nodeclipse 0.9.05)
NodeJs v 0.10.24
While this problem arise from time to time, the exact cause is not yet known. More feedback is needed.
(Once the problem goes away there is no way to reproduce healing; so the only way is to collect data in what environment, and how was solved)
Check Nodeclipse Online Help - Run page
There are 2 similar issues raised before.
#72, #176 was solved via Java upgrade to 1.7
#107 was solved via running node --debug-brk app.js from console first.
I have used Xdebug of Eclipse 3 months so far. And it runs really well.
However, I got a issue of Xdebug which is "DBGp Response Error: eval:=206 msg:error evaluating code" ( in Error Log).
After I investigate the error, I find that issue is:
206 - Error evaluating code (use from eval() (or perhaps property_get
for a full name get))"
This is reference from http://xdebug.org/docs-dbgp.php
I have taken 2 days for this issue, and nothing has changed.
Please help me solve this. Many thanks and strong appreciation for your help.
Try deleting your watches. You can read the story of how this worked for me in PHPStorm.
It is probably an eval that (your IDE) is doing wrong. Either the symbol does not exist, or the syntax is incorrect. I am not sure where your error_log comes from (do you mean xdebug.remote_log?), but that will also indicate what eval command was sent by the IDE. If you could amend your question with the exact log lines before and after your 206 error too, I can give a better clue. In general, this isn't much of an issue, as it's often not possible to eval specific code.
CentOS 7.7
PHP 7.2 and PHP 7.4
Apache and cli
Same symptoms " The strange in here is that it has run well, and suddenly can not debug with those things in error log" as #ndnguyen
Worked great. Then suddenly didn't work without explanation. I tried updating everything, several different config options, etc, etc.
Just that stupid little check box. FYI: The IDE is vscode.
I've been having problems getting debugging working with this setup for the past year. Supposedly, this is fixed in the latest versions, but I just can't seem to get it to work.
I'm using Mac OSX 10.8.5, Eclipse Keplar, PyDev 3.0.0, and Google App Engine 1.8.8.
I set up a run configuration to point to dev_appserver.py, running my project. That works fine. My python version is 2.7.3, and everything works in normal mode. When I try to run in debug mode, however, the app runs, the output specifically states:
pydev debugger: starting
pydev debugger: google app engine integration enabled
But after 20 seconds, the python process terminates with this error message:
'Launching New_configuration' has encountered a problem.
Timed out after 20.0 seconds while waiting for python script to connect.
Accept timed out
Execution works during those 20 seconds, because I can hit URLs connected to my app and see an effect. Breakpoints do not work, though. Is there something I need to do to get python to 'connect'?
Also - I should note that switching back to using 'old_dev_appserver.py' fixes this problem - debugging works fine when I do this. But, this is not really a solution, because I need to use the latest socket additions in the newer releases.
Thanks for any advice.
temporary solution:
edit the pydev_app_engine_debug_startup.py file:
eclipse/plugins/org.python.pydev_${VERSION}/pysrc/pydev_app_engine_debug_startup.py
change the line that reads:
if ':' not in config.version_id:
to:
if ':' not in config.version_id or config.version_id.startswith('default:'):
Looks like this is fixed in 1.9.0. I also had to make sure to update PyDev to the latest (3.3.3 at the time of this writing).
One more thing to add, and this one really tripped me up (on MacOS) -
In Eclipse, under Run -> Run Configurations, set the Main Module to point to this location:
/usr/local/google_appengine/dev_appserver.py
NOT the symlink:
/usr/local/bin/dev_appserver.py
Well, finally I found the answer. It is a Google's bug.
It has been reported and acknowledged.
see Debug is not working with PyDev again 1.8.8!
Update: Until the bug is fixed I am using PyCharm. The pro version has 30 days trial and works with GAE. It works like, well, a charm. I may even consider switching to it from Elicpse/PyDev
But I am used to JetBrains tools with Android Studio recnetly, and WebStorm and Resharper in the past, so I know to find my way around their IDEs. Your milage may and will vary.
Since the question, a new GAE has been released 1.8.9
Running GAE 1.8.9 with PyDev3.3.3 on Windows8.
This timeout issue still occurred.
When I entered the temporary workaround changes to pydev_app_engine_debug_startup described in the earlier answer, then the browser could not connect to the app.
However, on backing out this change, not only did the 'Launching New_configuration' not appear, but the breakpoints worked.
More info on the Python Tool Support for the App Engine Development Server is here
https://docs.google.com/document/d/1CCSaRiIWCLgbD3OwmuKsRoHHDfBffbROWyVWWL0ZXN4/edit