eSignal freezing during news today

[bf2]

Did anybody else experienced this today? Just before 10am ET my ES charts stopped refreshing and didn't come back to life until about five minutes and several restarts later.

I am running 10.5.

There was no problem with my net connection. It's a very resilient connection and it stayed up during the time. I was able to access any website I wanted, it's only eSignal that went down.

[jgr]

This is my 3rd attemp to reply. For some reason the forum times out.

I had the exact same problem. My system froze up and I had to use Ctrl-Alt-Del to kill it. By then the party was over.

jgr

[bf2]

Could someone from eSignal respond?

What is being done to resolve these problems in future? It's pretty bad if eSignal freezes every time Bernanke speaks.

[ScottJ]

While reports of freezing have almost disappeared, we have kept working on ways to maximize performance, especially during spikes in the market. There are many aspects to so-called "freezing" so we put together this KB article some time ago to help customers self-diagnose. Please review and see if anything stands out for your situation. Also be sure to be on the current 10.5 version.

Thanks.

[bf2]

Scott,
Thanks for your reply but I'm afraid that was more of a stock response.

Here's why:
1. My task manager never showed a heavy CPU load when the data froze. All other applications were working perfectly well. Hence my hardware was not a problem.

2. My internet connection stayed up all the time, so the fault was not there either.

3. My PC is fully up to date with the latest service packs, and I run eSignal 10.5.

I can only conclude that your data servers are not equipped to handle volume spikes, nor is the eSignal application.

So, what concrete steps are you taking to resolve these issues?

[jgr]

What has esignal identified as the functions that overload the system? Is it the number of graphic screens or the number of calculation for studies? Knowing that can help to better configure the system.

As far as the overall precautions in the KB, I have most of that under control. I do have 6 charts + two range charts (these seem to be a problem) and then $TICK + $TICKQ.

About 140 symbols.

FYI, I was running another appication (not as graphic intense as yours) on my laptop and it did not have this problem. It also was displaying the $TICK.

[bf2]

And I have only three charts showing just one instrument (ES), only one (the same) study on each chart (a moving average), configured to calculate on close.

That's all.

Hardly a configuration to bring the machine down.

[jgr]

I run ESET NOD32 for Vista 64. Are there any known issues with that? Of course, I am running 10.5 esignal.

[eSignal_AveryH]

NOD32 Anti-Virus has been known to cause issues with the Data Manager not being able to connect these two KB article 1415 and 1640 go into details of those issues.

[jgr]

I was running release 1657 of 10.5. I discovered that 1678 is out so I updated mine. Could that have been part of the freeze?

[ScottJ]

While it's always best to be on the latest, I don't recall any reported issues of performance degradation on that 1657 build. We're still reviewing the logs from that announcement on Friday but haven't found anything yet.

Thanks.

[jgr]

Today, the situation was worse. The entire esignal application just disappeared from the screens - no messages - no freeze - just gone. THIS happened at 10:00 AM when Consumer Confidence, FHFA Hoiuse Price and State Street Investor Confidence Index all were simultaneously announced.

Upon restarting, I had the freeze sitaution with grayed out screens (Vista 64). In the Windows Task Manager I saw several esignal components were "not responding". Among those was the "Alert" and "Paper trade". When I killed "Alert" all went back to normal.

SO, I am thinking that the issue has to do with the alerts that I have triggered in my custom efs. I have set up alerts on MACD cross and DI+ DI- cross situations.

I was running about 6 screens with MACD and DI+- alerts enabled. These both write messages to the "Triggered Alert List" and give audible alerts too.

Are there any known limits on how many of these things esignal can handle at a time? The complete crash usually means a serious issue with the application running out of its memory space or such.

I have the windows error report on this. Let me know if esignal wants for me to send it to someone.

[ScottJ]

Did you ever see the eSignal crash reporter appear? If so, I hope you sent that report so we could take a look. That's where you windows log would help even more.

Thanks.

[jgr]

There was no crash report this time. In the past that has come up. There is a Windows report of the crash with a lot of information. If you want I will email that to someone. Let me know. Vista is good about keeping these things.

[JayF]

jgr,

If ever eSignal gets into a locked up state, it automatically captures the state it's in for development purposes and will send this report to crash reporting servers to track. Thus, we likely already have the needed information.

Keep in mind that we have no major releases planned on this platform for this year. Our developers are nearly all focused on the new version of eSignal that runs on a completely new framework. Much of the existing code is being reworked, so many of these types of issues will likely disappear or at least be dramatically reduced.

That being said, it's still important to submit crash reports when you can. The more information we have, the better we can make the product.