Unable to Print to Printer Until Kills Java

[FinishingLine]

Strange one ... we have the same setup on several PCs (the PCs are all the same) with the same "normal" printer (on all but one) as well as a Zebra label printer

We have QZ loading just fine in the background and can print to the label printer as well with no issue, however, if we try printing to a "normal" printer, nothing comes out the printer (it appears like the job gets created but then is subsequently deleted)

Everything was working fine previously, though this issue just started on one computer and we also just changed the printers on the one that had a different printer so that everything was the same, but since doing that the issue seems to also now affect that one as well

Have tried both 2.1.1 (which we had originally) and also the latest 2.2.4-snapshot version

On the latest snapshot version of QZ, looking at the debug logs have spotted the below:

2024-04-23T11:42:10,082 [INFO] Connection opened from /0:0:0:0:0:0:0:1:63820 on socket port 8181
2024-04-23T11:42:10,082 [DEBUG] Message: {"certificate":"-----BEGIN CERTIFICATE-----\n....................\n-----END CERTIFICATE-----\n","timestamp":1713868930090,"uid":"..........","position":{"x":960,"y":540}}
2024-04-23T11:42:10,082 [WARN] Problem building certificate chain (normal if multiple CAs are in use)
2024-04-23T11:42:10,082 [DEBUG] Successfully chained certificate: CN=localhost, O=XXX (....................)
2024-04-23T11:42:10,082 [DEBUG] Adding XXX (localhost) to allowed list
2024-04-23T11:42:10,093 [DEBUG] Received new certificate from connection through 63820
2024-04-23T11:42:10,193 [DEBUG] Message: {"call":"getVersion","timestamp":1713868930095,"uid":".......","position":{"x":960,"y":540}}
2024-04-23T11:42:10,194 [DEBUG] Message: {"call":"printers.find","params":{"query":"ECOSYS P2040dn"},"signature":"........................","timestamp":1713868930095,"uid":"...........","position":{"x":960,"y":540}}
2024-04-23T11:42:10,194 [DEBUG] Searching for PrintService matching ECOSYS P2040dn
2024-04-23T11:42:10,194 [DEBUG] Matched default printer, skipping further search
2024-04-23T11:42:10,311 [DEBUG] Message: .................
2024-04-23T11:42:10,327 [DEBUG] Using qz.printer.action.PrintPDF to print
2024-04-23T11:42:10,327 [DEBUG] Searching for PrintService matching ECOSYS P2040dn
2024-04-23T11:42:10,327 [DEBUG] Matched default printer, skipping further search
2024-04-23T11:42:10,327 [WARN] Raw option "altPrinting" is deprecated.  Please use "forceRaw" instead.
2024-04-23T11:42:13,311 [INFO] Connection closed: 1001 - null
2024-04-23T11:42:13,311 [INFO] Closing all communication channels for localhost
2024-04-23T11:42:14,084 [DEBUG] Parsed 1 files for printing
2024-04-23T11:42:14,178 [INFO] Starting printing (1 copies)
2024-04-23T11:42:14,287 [DEBUG] Paper area: 0,0:597,842
2024-04-23T11:42:14,413 [DEBUG] Paper area: 0,0:597,842
2024-04-23T11:42:14,413 [DEBUG] Paper area: 0,0:597,842
2024-04-23T11:42:14,491 [INFO] Printing complete
**2024-04-23T11:42:14,506 [ERROR] Could not send message
java.io.IOException: java.nio.channels.ClosedChannelException
	at org.eclipse.jetty.util.FutureCallback.block(FutureCallback.java:163) ~[qz-tray.jar:?]
	at org.eclipse.jetty.util.FutureCallback.block(FutureCallback.java:139) ~[qz-tray.jar:?]
	at org.eclipse.jetty.websocket.common.JettyWebSocketRemoteEndpoint.sendBlocking(JettyWebSocketRemoteEndpoint.java:191) ~[qz-tray.jar:?]
	at org.eclipse.jetty.websocket.common.JettyWebSocketRemoteEndpoint.sendString(JettyWebSocketRemoteEndpoint.java:52) ~[qz-tray.jar:?]
	at qz.ws.PrintSocketClient.send(PrintSocketClient.java:793) ~[qz-tray.jar:?]
	at qz.ws.PrintSocketClient.sendResult(PrintSocketClient.java:724) ~[qz-tray.jar:?]
	at qz.utils.PrintingUtilities.processPrintRequest(PrintingUtilities.java:215) ~[qz-tray.jar:?]
	at qz.ws.PrintSocketClient.processMessage(PrintSocketClient.java:315) ~[qz-tray.jar:?]
	at qz.ws.PrintSocketClient.lambda$onMessage$0(PrintSocketClient.java:169) ~[qz-tray.jar:?]
	at java.base/java.lang.Thread.run(Unknown Source) [?:?]
Caused by: java.nio.channels.ClosedChannelException
	at org.eclipse.jetty.websocket.core.internal.WebSocketSessionState.onOutgoingFrame(WebSocketSessionState.java:184) ~[qz-tray.jar:?]
	at org.eclipse.jetty.websocket.core.internal.WebSocketCoreSession.sendFrame(WebSocketCoreSession.java:514) ~[qz-tray.jar:?]
	at org.eclipse.jetty.websocket.common.JettyWebSocketRemoteEndpoint.sendBlocking(JettyWebSocketRemoteEndpoint.java:190) ~[qz-tray.jar:?]
	... 7 more**

I'm wondering if the exception is the reason?

When trying a different brand printer, the same happens, though this particular printer starts whirling like its about to print, but then the job gets cancelled and it just stops

The strange thing is; once we kill Java via Task Manager and re-open QZ, everything starts working fine for the "normal" printer (and the label printer continues to works as it did before)

Originally wondered if this was something to do with a Windows Update, but if that then the other machines should have the same issue

[tresf]

Thanks for the detailed report.

My initial observations...

• I would expect something corrupted with the printer spooler. This is usually hard to track down, but "normal printer" is a bit subjective, I'd make sure it doesn't happen with Microsoft Print to PDF, since this uses a Microsoft-provided driver and should be less likely to cause this problem. Operating Systems like Mac actually have a way to fully reset the printing subsystem, but as far as I know, Windows does not, so truly deleting all printers and drivers to troubleshoot this is a painstaking process.
• I'm sure you've tried this, but completely removing and grabbing a different Kyocera driver is always a good idea. For bad Chinese perinters, I've noticed that newer isn't always better. Kyocera, on the other hand, is a well respected brand, so I would expect the latest driver to try to address any crashes or conflicts with the OS, assuming they've been reported.
• If you were to start QZ Tray from terminal (C:\Program Files\QZ Tray\qz-tray-console.exe) you might actually get a stacktrace into Java's C++ code. This is seldom helpful, but would at least help pinpoint the issue as a hard JVM crash that can't be remedied with any Java code.
• If by (rare) chance the software stays alive and the exception is just some weird uncaught exception, we'd be happy to try to patch that section of code to gather more details.
• If by chance it's an endless loop, you should see a CPU spike. We might be able to debug that as well.
• As the above bullets alude, the exception doesn't really help... I believe the exception to be the symptom of a race condition between the software crashing and the websocket trying to report the crash, hence it having nothing to do with printing. I've seen this same message before when I use Task Manager to kill the java[w].exe process.
• We've also seen some rare edge-case issues where the software crashes are due to external factors (usually graphic related), so it's always possible that the problem is specific to a particular hardware profile. For example, we once had an issue where QZ Tray would crash using a dongle on a Mac. In that case, we reached out to the dongle manufacturer with a small test case and they fixed the firmware, correcting the problem.

[tresf]

Oh... and this one is just a "have you turned your computer off and back on again" recommendation, but nothing has caused more grief in my modern troubleshooting life than pending windows updates that haven't been applied and worn bits on an SSD hard drive, so as a general rule of thumb, before I go insane, I do a chkdsk /f /r on any problematic computer.

[FinishingLine]

@tresf Many thanks for the above

For clarity, this issue happens at bootup (and we've rebooted a lot of times when trying to figure this one out), once Java is killed, everything is fine. The label printer also continues working after the exception, without needing to restart/reload QZ

Have also tried doing a full factory reset on the PC, but the issue didn't disappear

We have tried using the IPP driver but will also look to see if we can use a different Kyocera driver as well (and will try to completely remove the existing one too)

Will also give print to PDF a try

[tresf]

For clarity, this issue happens at bootup

Can you describe the symptom in more detail? I'm not sure that I understand. The original bug report show the websocket closure occuring fractions of a second after a print occurs. I assume when you say "Happens at bootup", what you're describing is:

1. When QZ Tray starts for the first time (specifically, when started through the routine startup process), it misbehaves, but only after a PIXEL print is performed (I assume this due to the label printer being OK, but please correct this statement if it's wrong).
2. When QZ Tray starts for the second and subsequent times (after being manually killed) it behaves properly

Is this correct?

Furthermore, if you wait long enough, does it eventually crash on its own? Are there any other symptoms that would suggest that QZ Tray is still performing any type of operation (CPU, ram, etc?). You've peaked my interest.

We did have an oddball issue where 3rd party apps caused troubles paired with QZ Tray (it has a group policy to do JVM enforcement and we were being punished). In the case of one, it was using some advanced techniques to affect DLL search and load order. If something similar is happening on your machine (an incorrect DLL is being confused at startup) it may explain the odd behavior. This is just a shot in the dark.

Another shot in the dark is to see if something is stealing port 8181. We try to fallback on 8282 if 8181 is already taken, but perhaps something is breaking in this process.

[FinishingLine]

@tresf just to give you an update, we played around with various drivers and have found one that doesn't seem to cause the issue :)

Microsoft Print to PDF also seemed to work when we changed the printer, so perhaps it was a driver issue that oddly only affected 2 out of 5 machines (not sure if possibly as we removed and added the printer back on these two, but setup was exactly the same across all 5)

Can you describe the symptom in more detail? I'm not sure that I understand. The original bug report show the websocket closure occuring fractions of a second after a print occurs. I assume when you say "Happens at bootup", what you're describing is:

1. When QZ Tray starts for the first time (specifically, when started through the routine startup process), it misbehaves, but only after a PIXEL print is performed (I assume this due to the label printer being OK, but please correct this statement if it's wrong).
2. When QZ Tray starts for the second and subsequent times (after being manually killed) it behaves properly

Is this correct?

Yes, this is what was happening (loading on startup) in both versions

Furthermore, if you wait long enough, does it eventually crash on its own? Are there any other symptoms that would suggest that QZ Tray is still performing any type of operation (CPU, ram, etc?). You've peaked my interest.

It didn't crash per se, it just refused to print (by the print job being deleted without us doing anything) to the printer until Java was killed and QZ was re-opened; printing to a Zebra label printer via QZ tray worked just fine

Didn't spot any obvious CPU/RAM spikes when attempting to print

[tresf]

@FinishingLine thanks for the detailed explanation. Is this still happening and something you believe requires a remote session to escalate, or do you plan on considering this an edge-case issue that we can close and revist at a later time?