Support Board
Date/Time: Wed, 07 Oct 2026 12:42:52 +0000
v2922 & v2947: study calculation or replay data loading silently stops chartbook-wide during Accurate back test replay; Reca
View Count: 286
| [2026-09-09 16:18:44] |
| User313107 - Posts: 4 |
|
Sierra Chart 2922 64-bit (rev 57220) and 2947 64-bit (rev 58042), Windows 11, Not Connected (all networking blocked by firewall during back tests by design). Chartbook: 4 charts of one symbol (60 min, 5 min, 2 min, 1 min), custom ACSIL studies with cross-chart references (charts #1 and #2 feed chart #4), Replay Mode = Accurate Trading System Back Test Mode, All Charts in Chartbook, 100X, Skip Empty Periods off, Chart Update Interval 400 ms, Multiple Chart Back Test Update Interval 200 ms. Reproduced with "Use OS Timer For Charts" = Yes and = No, and on both versions (2947 was a clean controlled update over the same chartbook/data/config; we have since rolled back to 2922 because the behavior also occurs there and comparability of results was lost). Instrumentation: one ACSIL study ("ACS_STUDY_A" in the logs; manual loop, sc.UpdateAlways = 1) logs a heartbeat every 1000 invocations and every 100 bars via sc.AddMessageToLog. Normal cadence during replay data: one line every ~31 s (~35-45 calls/s); during a no-data period one every ~220 s with the bar index static. Custom study/DLL/chartbook names in the logs are neutralized (ACS_STUDY_A, CUSTOM_DLL_A...E, CHARTBOOK_A) - consistently, so the sequence of events is fully traceable. Occurrence 1 - 2922, OS Timer = Yes (log: FIXRUN1_LOG_redacted.log): at 46 min into the replay (replay time ~ 08-03 23:10), every study on every chart stopped being calculated at the same moment, while the replay clock kept advancing, new bars kept appending on all charts, redraw/menus/scrolling worked, and the process later exited normally. CPU ~0.4 %, private memory ~71 MB, flat. Last heartbeat calls=40000, bars=34930 at 16:49:35; for the following 52 minutes not a single heartbeat (the scsf function was never invoked again) while the 1-min chart gained ~3,000 bars. Nothing was logged at the moment it stopped. Not restored by: Pause then Play (replay resumed, studies still dead); Chart - Recalculate: no "Reloading chart" followed, and shortly after, "Volume By Price ... SubGraphData ... Array is unallocated" (arrays freed, recalculation never ran); Study Settings - Apply; Replay Stop: "Reloading chart" was logged for all 4 charts but "Chart data loading complete" never followed (4+ min, charts blank), "Back test processing: NumberOfTimesCalculateCalled (Cumulative) = 0" per chart. File - Exit normal. Occurrence 2 - 2922, OS Timer = No (log: FIXRUN2_LOG_redacted.log; same chartbook/data/start): at 62 min in (replay time 08-05 00:27:41) a different stage froze: replay data loading. LastLoadedDateTime stuck permanently (the .scid is continuous at that point) while study calculation kept being invoked at the no-data cadence. Pause/Play no effect; a Replay "Jump" was accepted and logged but never applied; rewind: "Reloading chart" never completed; Stop produced no log line; Exit normal. Occurrences 3-4 - 2947 (log: AB_RUN3_2947_LOG_redacted.log; same chartbook/data/config after a clean controlled update): (a) With Replay Start 17:59 (inside data), study calculation was invoked once and then never again from 38 s after Play (bars static). After Stop, the charts logged loading complete, but the full recalculation that chart #1/#2 had tagged for chart #4 never ran: "Base Graph ... Array is unallocated", no heartbeat, chart #4 stayed blank until Exit. (b) Separately: starting the replay at a date-time inside a no-data period (Start 08-02 00:00, a Saturday) leaves the replay position at the prior Friday close (07-31 16:59:54) indefinitely - 25 min observed, studies alive, Stop/reload normal. On 2922 the same start advances immediately. We found nothing about this change in What's New 2923-2947 - is it intended? Message Logs were searched for "Error posting message in c_AsyncWindow::PostMessage", "Not enough quota", "PostMessage", "message queue" and related text across the three sessions above plus four earlier ones (740-4,714 lines each): zero occurrences. The end state matches thread 106616 (v2891/v2927, live data): all studies stop, data path alive, threads idle, CPU ~idle, restart required - here under replay, on 2922 and 2947, with either timer setting, and in two variants (calculation stops while data advances; data loading stops while calculation continues). Questions: 1. Is this a known condition, and do the 2931 timer/message-queue changes address it? (We still reproduce on 2947.) 2. Which dump/log method do you want for a replay-time occurrence, given nothing is logged at the freeze moment? We can reproduce within ~1-2 hours of replay. 3. Is the replay-start-inside-a-no-data-period behavior on 2947 (item (b)) intended? 4. Is anything known to interact between cross-chart study references (tagged full recalculation under controlled-order updating) and this condition? Note: the three Message Logs and five screenshots referenced above will be attached immediately after posting (attachment limit permitting). Date Time Of Last Edit: 2026-09-09 16:37:42
|
Private File Private File Private File |
| [2026-09-09 16:40:10] |
| User313107 - Posts: 4 |
|
Screenshots (part 1 of 2), referenced in the first post. Custom study/chartbook names are masked in the images, consistent with the neutralized names in the logs. - SS_run1_blank_charts_after_stop.png: occurrence 1 (2922, OS Timer Yes) - charts blank after Replay Stop; 'Reloading chart' logged for all 4 charts but 'Chart data loading complete' never followed. - SS_run2_blank_after_jump.png: occurrence 2 (2922, OS Timer No) - blank charts after the Jump that was accepted and logged but never applied. - SS_run2_UseOSTimer_No.png: setting confirmation for occurrence 2. Part 2 with the two 2947 screenshots follows. |
 SS_run1_blank_charts_after_stop.png / V - Attached On 2026-09-09 16:38:51 UTC - Size: 36.71 KB - 42 views SS_run2_blank_after_jump.png / V - Attached On 2026-09-09 16:39:22 UTC - Size: 54.52 KB - 44 views SS_run2_UseOSTimer_No.png / V - Attached On 2026-09-09 16:39:55 UTC - Size: 8.53 KB - 38 views |
| [2026-09-09 16:41:52] |
| User313107 - Posts: 4 |
|
Screenshots (part 2 of 2), from the 2947 controlled-update session: - SS_2947_replay_not_advancing.png: occurrence (b) - Replay Start 2026-08-02 00:00 (inside a no-data period); the replay position stays at 07-31 16:59:54 (visible in the top bar) while the dialog shows the requested start date; observed for 25 min. - SS_2947_blank_chart4_after_stop.png: occurrence (a) - chart #4 blank after Replay Stop; the tagged full recalculation never ran (C: 0.0, no heartbeat until Exit). This completes the evidence set: 3 Message Logs on the first post, 5 screenshots across these two replies. We can provide further captures or run a specific procedure on request. |
| SS_2947_replay_not_advancing.png / V - Attached On 2026-09-09 16:41:08 UTC - Size: 98.78 KB - 37 views SS_2947_blank_chart4_after_stop.png / V - Attached On 2026-09-09 16:41:37 UTC - Size: 59.37 KB - 37 views |
| [2026-09-09 16:48:40] |
| John - SC Support - Posts: 48127 |
|
If you try this without the custom studies, do you have the same issue? Can you reproduce this without the custom studies? An easy way to do this is to use the "Safe Mode" option when logging into Sierra Chart. Refer to the following: Sierra Chart - Software Login Instructions: Using Safe Mode For the most reliable, advanced, and zero cost futures order routing, use the Teton service: Sierra Chart Teton Futures Order Routing |
| [2026-09-11 23:31:55] |
| User313107 - Posts: 4 |
|
Update - isolation test per your suggestion (Safe Mode, no custom studies). Setup: a disposable copy of the installation (2922 64-bit, same data files, a copy of the chartbook, network disabled at the OS). Same replay recipe as the earlier reports: Accurate Trading System Back Test, All Charts in Chartbook, start 2026-08-02 00:00:00, speed 100X, Skip Empty Periods off, Processing Step 1 second. Two runs on 2026-09-11, same machine, fresh program launch for each. Note the two runs were stopped at different points (details below), so the Stop conditions are not identical between them. RUN A - normal login (custom studies loaded): - Replay ran ~2.0 hours real time (2026-08-02 -> ~2026-08-08) with NO mid-run stall; it passed both replay points where the earlier stalls happened. - Pressed Stop at 14:22:11 (chart local time): "Reloading chart" was logged for all 4 charts, then nothing further appeared in the message log until program exit at 14:29:55 - 7 min 44 s with no "Chart data loading complete". Charts were blank in a screenshot taken 16 s after Stop and still blank just before exit. File >> Exit closed normally. RUN B - Safe Mode login (message log shows "Safe Mode enabled"; zero custom study / DLL lines in the log for the whole session; no custom studies computing): - Identical replay recipe. This run was left running longer, ~3.5 hours real time (2026-08-02 -> 2026-08-09 19:41), with NO mid-run stall. - Pressed Stop at 18:14:23: same failure - "Reloading chart" x4 logged, then nothing further in the message log until exit at 18:22:40 - 8 min 17 s with no "Chart data loading complete". Charts blank in a screenshot at 18:22:06 (7 min 43 s after Stop). Exit normal. Answering your questions: 1) The post-Stop dead-reload reproduced in both Normal and Safe Mode - i.e. this symptom occurred without the custom studies loaded in this test. That rules out the custom studies as a required condition for THIS symptom in this trial; it does not by itself confirm the root cause. 2) The mid-run silent stall from the earlier reports (study calculation stopping at ~46 min in one run; the replay data loader freezing at ~62 min in another) did NOT occur in either run this session. It is intermittent, so we cannot yet say whether that component reproduces without custom studies. For completeness on the reload-failure history by trigger: after Stop - run 1 (2922), run 3c (2947), and now both of these runs (2922, with and without custom studies); after a backward Jump - run 2 (2922). One observation from these two runs (this session only): with this start time (Sunday 00:00, Skip Empty Periods off) the replay position stayed at the previous close (07-31 16:59:54) for ~28 minutes of real time before data began, then proceeded normally, in both runs. Our earlier report of the replay never leaving the start position on 2947 was stopped at 25 minutes; whether that was the same waiting behavior or a real failure is unresolved - we have not re-verified it. Message logs and screenshots from both runs are available - happy to attach if useful. Date Time Of Last Edit: 2026-09-11 23:36:44
|
| [2026-09-14 13:55:56] |
| John - SC Support - Posts: 48127 |
|
Let's concontrate on what you are stating is the "Stop dead-reload". When you Stop the replay, you are saying that the charts are not reloading and getting back to their normal state showing all the data to the current time? Are you definitely selecting the "Stop" button on the Replay window? What exactly does the chart show in this case? Please get us an image of what you see after you stop the replay along with another image of what the chart should look like normally. Follow these instructions to get us images: Support Board Posting Information: Attaching or Posting a Screenshot/Image (Optional) And let us know what version of the software you are running when you do this. Preferably one of the latest releases. For the most reliable, advanced, and zero cost futures order routing, use the Teton service: Sierra Chart Teton Futures Order Routing |
To post a message in this thread, you need to log in with your Sierra Chart account: