Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
It's true that you can get issues with workspaces, but again this most likely comes back to indicators. There's plenty of ways to screw things up by not managing state, serialization, public variables, and edge conditions properly. There can be bugs in the indicator that don't show up until you reload the workspace or template. You run into an edge condition that your code doesn't handle properly, and then that error gets serialized and causes problems the next day.
My suggestion for the NinjaTrader devs is to focus on some changes that will help us track down issues like this more easily. Here's a few ideas:
When indicators break I usually just get a short error message to the output window / log. This is rarely enough information to help anyone determine even what indicator the error is coming from. I'll usually try to debug with visual studio so I can fix it, but most users aren't going to take it that far. I'll usually need a stack trace, and information about which chart it's attached to in order to track it down.
I've had issues where it crashes while loading up the workspaces. Deleting my workspace.xml file solves the issue, but doesn't allow me to find out why it was crashing. The ability to pause NinjaTrader during startup so that I can attach a debugger to the application would be useful.
I have a few bugs that are difficult to recreate. I'll only run into a day that causes the error once a month, and when I do I don't have time to debug things. When I start up the next day it's gone. Some way to save the state of everything so I can debug later would be nice.
It would also be nice to have a way to specifically recreate problem datasets. An instrument with hand crafted historical data that I know has caused problems with my indicator in the past. I could then use that instrument as a sort of regression test.
A way for the author to provide within the script contact and update information. If an error with an indicator is detected the users can send bug reports to the author. If the author has made an update it can notify the users.
If you're having these sort of workspace issues my suggestion is to list out every indicator you're using in your workspace. Perhaps we can find issues in the indicators you are using.
Thanks for all of your input. Has any of you experience these problems?
1) My workspaces go "blank" on a regular basis, that is the file is not deleted but the content is empty. This usually happens after an automatic backup which is every Friday.
2) when I'm trying to load/save a template, the menu bar on each chart flashes erratically. I'm told that there's a bug with one of my code, as instructed by NT I start fresh and tested each indicator one by one, I can't replicate the problem. But when I run these indicators together, the problem comes back after awhile. NT solution is to reinstall the program again. I really don't want to do this as I've spent so much time fixing things on NT rather than trading. I'm using V8.05.
If anyone has similar experiences or know what the problem/fix could be, please let me know. Thanks!
Hi Luigi.... I have had an occasional flashing of the chart, but not the other probs. I do not use the Ninja backup. I back up my entire drive weekly.
My routine seems to work pretty well for keeping NT8 in check..... I posted it a few days ago in this thread...
Daily....
Before I start, I open the Windows Task Manager and sort by memory use.
Before I start NT 8, I clear NT 8 cache and in db clear cache and the contents of day, min, and tick as well as _Workspaces (bcNT8Cleanup.bat)
After I start NT 8, I let the NT8 CPU usage go down under 10% and then I connect to broker.
Memory should be in the 700Mb area, depending on your charts and indis.
I keep an eye on memory use and if it creeps up to 1.5GB area, I go through the NT shutdown and restart procedure.. Changing indi parameters may start a memory build.
Every Week, delete titled workspaces and rebuild them. I don’t know why, but they seem to contribute to memory creep. To rebuild, start a fresh chart and apply the chart templates. So far, the chart templates do not add to memory creep.
With all of the probs that I am reading about and experiencing, it seems that NT either needs to set up a service to produce compliant indis or have a way of validating the indies that the users build. The swamp seems to be getting a lot busier.
Thanks Blash.... I have asked NT to supply some tool that would allow them, or users to track down what is consuming the memory... even as basic as the Windows Task manager, but with enough detail to watch each window and perhaps each process within a window within NT. I asked several months ago and have not seen anything yet. Maybe if we all post enough, they will get the message. The platform has great potential, but the frustrations may outway the benefits at some point. NT Support seems to accept mediocre platform performance... or their hands are tied and they have no solution. I agree with you... it is a development problem.
Thanks TWDsje, to me, it seems that we are chasing a ghost, but that just means that I dont understand yet. Using my data and cache clearing routing daily seems to keep things in check, but I am continuing on the path of trying to find if my indies are at the root of the problem.
I know that there is no data flowing on the weekend, but I have had memory creep problems on weekends in the past. I loaded 10 instances of each of my indies on a chart, one by one, watching for odd behavior. I changed parameters around a lot and recalculated often and I could not find any issues on this experiment. Perhaps the latest version has resolved some of those issues. I am running 8.0.5.1 64-bit.
Maybe this week, I will try not using my clear routine and see if the memory issues have been resolved.
We should all keep piping in and see if we can help each other in any way. I know that SIM22 really helped me a lot, and Zondor has given me some info to digest.
My comments are based mostly on things that I've seen in my own indicators. So for instance I had an issue where certain properties were public instead of private, and this was causing data to get serialized that shouldn't have been. Then that data was overwriting what was set up in my initialization, and caused all sorts of bizarre issues. You can imagine how if you save your workspace every day that this data could get reinitialized every time, and cause an issue later on down the line.
But there also seems to be some sort of edge case in indicators that use OnMarketData that I simply can't track down. It happened today in ZN's ETH with Rithmic as the datafeed, but of course I couldn't investigate because I was trading. Today it even messed up my Jigsaw DOM. Changing my chart to the RTH fixed the error.
I have noticed that if I set an Alert to Share a message to Twitter with Screenshot type: Tab, having a chart window with many tabs, moving to a different tab, leaving my office and an hour later the Alert is hit ....the chart pic is from the tab I was last viewing and not from the chart Tab containing the Alert.
This must be an error correct? The point is to have a chart picture from where the Alert was generated ...not some other chart.
Say I have a chart window with 5 Tabs and one Alert per chart..... I have to end on one Tab I can't be viewing all of them. So for each Alert the same chart pic would be sent?????
Thank you.
Ron
...My calamity is My providence, outwardly it is fire and vengeance, but inwardly it is light and mercy...
The steed of this Valley is pain; and if there be no pain this journey will never end.
Buy Low And Sell High (read left to right or right to left....lol)
I have seen these kind of things happen with some of my custom indicators. I think there is a major bug or architectural flaw in NT8 relating to tabs. In my case when using tabbed charts some members and methods become "shared" among the tabs rather that being specific to its own tab chart.
I have not looked into it further, my solution is to place each chart in its own window avoiding tabs altogether.
