Windows IRPStackSize increased to pass test-debug-port-from-cmdline.js #2094

rvagg · 2015-07-02T08:55:16Z

@nodejs/platform-windows

test-debug-port-from-cmdline.js has been failing repeatedly on 2008 like so:

not ok 136 - test-debug-port-from-cmdline.js
#c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20
#    process._debugProcess(child.pid);
#            ^
#Error: Not enough storage is available to process this command.
#    at Error (native)
#    at ChildProcess.onChildMsg (c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20:13)
#    at emitTwo (events.js:87:13)
#    at ChildProcess.emit (events.js:172:7)
#    at handleMessage (internal/child_process.js:632:10)
#    at Pipe.channel.onread (internal/child_process.js:421:11)

So I followed the instructions @ https://support.microsoft.com/en-us/kb/106167 and increased IRPStackSize to 9 and now it appears to be passing.

This doesn't seem like an appropriate course of action because we can't advise developers to do this in order to get the test suite to pass. But since I have no idea why this is even required I'm not sure of a better way forward or how to suggest a possible fix.

The text was updated successfully, but these errors were encountered:

benjamingr · 2015-07-02T09:04:41Z

Isn't 9 well below the default value?

On Thu, Jul 2, 2015 at 11:55 AM, Rod Vagg [email protected] wrote:

@nodejs/platform-windows
https://github.com/orgs/nodejs/teams/platform-windows

test-debug-port-from-cmdline.js has been failing repeatedly on 2008 like
so:

not ok 136 - test-debug-port-from-cmdline.js
#c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20

process._debugProcess(child.pid);

^

#Error: Not enough storage is available to process this command.

at Error (native)

at ChildProcess.onChildMsg (c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20:13)

at emitTwo (events.js:87:13)

at ChildProcess.emit (events.js:172:7)

at handleMessage (internal/child_process.js:632:10)

at Pipe.channel.onread (internal/child_process.js:421:11)

So I followed the instructions @
https://support.microsoft.com/en-us/kb/106167 and increased IRPStackSize
to 9 and now it appears to be passing.

This doesn't seem like an appropriate course of action because we can't
advise developers to do this in order to get the test suite to pass. But
since I have no idea why this is even required I'm not sure of a better way
forward or how to suggest a possible fix.

—
Reply to this email directly or view it on GitHub
#2094.

rvagg · 2015-07-02T09:11:22Z

I believe 4 is the default when the entry doesn't exist

benjamingr · 2015-07-02T09:13:33Z

The default value of the IRPStackSize parameter is 15. The range is from 11 (0xb hexadecimal) through 50 (0x32 hexadecimal).

https://support.microsoft.com/en-us/kb/285089

rvagg · 2015-07-02T09:15:59Z

I see your KB article and raise you mine:

Note Values may range from 0x1 to 0xC. These values are equivalent to 1 to 12 in decimal notation.

If the IRPStackSize registry entry is not present, the computer uses a default value of 0x4.

https://support.microsoft.com/en-us/kb/106167

But then again, it has NT 3.51 at the bottom of the page!

benjamingr · 2015-07-02T09:20:18Z

But then again, it has NT 3.51 at the bottom of the page!

Haha, yeah, it even links to a different one from NT 4, I wonder what running io.js on any of those would look like. Anyway, I'll try a fresh VM and check.

benjamingr · 2015-07-02T09:45:03Z

Ok, opened a new Windows Server 2008R2 VM on Azure - it is not defined:

Would it help if I cloned io.js on it and ran the tests?

rvagg · 2015-07-02T09:48:07Z

Anyone running tests on windows helps! we have too few eyes at the developer level on core code & tests on windows. See https://github.com/nodejs/build/tree/master/setup/windows if you want detailed docs of how we do it.

benjamingr · 2015-07-02T09:52:50Z

Thanks, I'm already familiar with the procedure :) Any specific branch?

rvagg · 2015-07-02T09:54:24Z

master is where we're seeing it, also I've upped the 2008 servers to 24 (decimal) now

benjamingr · 2015-07-02T10:02:08Z

Hmm, on Windows 2008 we can't install VS2013 Express since it requires a newer version of Windows. I'll try with VS2012

rvagg · 2015-07-02T10:15:01Z

vs2012 isn't going to work, you need Microsoft Visual Studio Express 2013 for Windows Desktop or a prerelease of 2015, we're running on 2008R2 just like your Azure machine, perhaps it's the fact that you're trying Express that's the problem.

mathiask88 · 2015-07-02T10:44:27Z

There is a difference between "Express for windows" and "Express for windows desktop". The last one is compatible with Windows Server 2008 R2 SP1 (x64).

orangemocha · 2015-07-02T12:40:22Z

You can also use Visual Studio Community Edition. It practically supersedes Express: https://www.visualstudio.com/en-us/products/visual-studio-community-vs.aspx

I just learned about IRPStackSize, but it seems an issue with the specific driver stack on the machine, not a general issue. See this post on the subject: http://forum.sysinternals.com/when-irpstacksize-can-cause-a-problem_topic22059_post116568.html#116568

silverwind · 2015-07-04T11:48:08Z

Agree with @orangemocha, the stack in question here seems to be the machine's own driver stack, which is clearly out of control for us, so there's not much we can do, except maybe parsing for that particular error message.

silverwind · 2015-07-06T22:37:04Z

Same test now seems to fail with an cryptic 'Access is denied', ugh

not ok 136 - test-debug-port-from-cmdline.js
# c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20
# process._debugProcess(child.pid);
# ^
# Error: Access is denied.
# at Error (native)
# at ChildProcess.onChildMsg (c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20:13)
# at emitTwo (events.js:87:13)
# at ChildProcess.emit (events.js:172:7)
# at handleMessage (internal/child_process.js:632:10)
# at Pipe.channel.onread (internal/child_process.js:421:11)

https://jenkins-iojs.nodesource.com/job/iojs+pr+win/79/nodes=win2008r2/tapTestReport/test.tap-136/

rvagg · 2015-07-10T00:09:17Z

I'd appreciate eyes on this: #2149

Some driver stacks on win32 will trigger an occasional error on test-debug-port-from-cmdline.js, reporting: Not enough storage is available to process this command Workaround is to check for that error and ignore the test entirely. Closes: nodejs#2094

This test was failing because the spawned process was terminated before anything could be done, by calling child.stdin.end. With this change, the child's stdin is no longer closed. When the stdin is not a tty, io.js waits for the whole input before starting, so the child must be run with --interactive to process the command sent by the parent. The child is killed explicitly by the parent before it exits. This test was failing silently because the asserts were not called if nothing was received from the child. This fix moves assertOutputLines to always run on exit. Fixes nodejs#2094 Fixes nodejs#2177

This test was failing because the spawned process was terminated before anything could be done, by calling child.stdin.end. With this change, the child's stdin is no longer closed. When the stdin is not a tty, io.js waits for the whole input before starting, so the child must be run with --interactive to process the command sent by the parent. The child is killed explicitly by the parent before it exits. This test was failing silently because the asserts were not called if nothing was received from the child. This fix moves assertOutputLines to always run on exit. Fixes: #2177 Refs: #2094 PR-URL: #2186 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Rod Vagg <[email protected]> Reviewed-By: Johan Bergström <[email protected]> Reviewed-By: Alexis Campailla <[email protected]>

cjihrig · 2015-07-16T02:01:11Z

Going to close this. The problem is hopefully fixed in 2b4b600.

This change is a backport of 2b4b600 from io.js. Original commit message: This test was failing because the spawned process was terminated before anything could be done, by calling child.stdin.end. With this change, the child's stdin is no longer closed. When the stdin is not a tty, io.js waits for the whole input before starting, so the child must be run with --interactive to process the command sent by the parent. The child is killed explicitly by the parent before it exits. This test was failing silently because the asserts were not called if nothing was received from the child. This fix moves assertOutputLines to always run on exit. Fixes: nodejs/node#2177 Refs: nodejs/node#2094 PR-URL: nodejs/node#2186 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Rod Vagg <[email protected]> Reviewed-By: Johan Bergström <[email protected]> Reviewed-By: Alexis Campailla <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> PR-URL: #25748

This change is a backport of 2b4b600 from io.js. Original commit message: This test was failing because the spawned process was terminated before anything could be done, by calling child.stdin.end. With this change, the child's stdin is no longer closed. When the stdin is not a tty, io.js waits for the whole input before starting, so the child must be run with --interactive to process the command sent by the parent. The child is killed explicitly by the parent before it exits. This test was failing silently because the asserts were not called if nothing was received from the child. This fix moves assertOutputLines to always run on exit. Fixes: nodejs/node#2177 Refs: nodejs/node#2094 PR-URL: nodejs/node#2186 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Rod Vagg <[email protected]> Reviewed-By: Johan Bergström <[email protected]> Reviewed-By: Alexis Campailla <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> PR-URL: nodejs#25748

mscdex added windows Issues and PRs related to the Windows platform. test Issues and PRs related to the tests. labels Jul 2, 2015

silverwind mentioned this issue Jul 14, 2015

test-debug-port-from-cmdline odd behaviour #2177

Closed

joaocgreis mentioned this issue Jul 15, 2015

test: fix test-debug-port-from-cmdline #2186

Closed

cjihrig closed this as completed Jul 16, 2015

joaocgreis mentioned this issue Jul 22, 2015

test: fix test-debug-port-from-cmdline nodejs/node-v0.x-archive#25748

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Windows IRPStackSize increased to pass test-debug-port-from-cmdline.js #2094

Windows IRPStackSize increased to pass test-debug-port-from-cmdline.js #2094

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

process._debugProcess(child.pid);

^

at Error (native)

at ChildProcess.onChildMsg (c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20:13)

at emitTwo (events.js:87:13)

at ChildProcess.emit (events.js:172:7)

at handleMessage (internal/child_process.js:632:10)

at Pipe.channel.onread (internal/child_process.js:421:11)

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

mathiask88 commented Jul 2, 2015

orangemocha commented Jul 2, 2015

silverwind commented Jul 4, 2015

silverwind commented Jul 6, 2015

rvagg commented Jul 10, 2015

cjihrig commented Jul 16, 2015

Windows IRPStackSize increased to pass test-debug-port-from-cmdline.js #2094

Windows IRPStackSize increased to pass test-debug-port-from-cmdline.js #2094

Comments

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

process._debugProcess(child.pid);

^

at Error (native)

at ChildProcess.onChildMsg (c:\workspace\iojs+pr+win\nodes\win2008r2\test\parallel\test-debug-port-from-cmdline.js:20:13)

at emitTwo (events.js:87:13)

at ChildProcess.emit (events.js:172:7)

at handleMessage (internal/child_process.js:632:10)

at Pipe.channel.onread (internal/child_process.js:421:11)

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

benjamingr commented Jul 2, 2015

rvagg commented Jul 2, 2015

mathiask88 commented Jul 2, 2015

orangemocha commented Jul 2, 2015

silverwind commented Jul 4, 2015

silverwind commented Jul 6, 2015

rvagg commented Jul 10, 2015

cjihrig commented Jul 16, 2015