Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hello,
we recently upgraded from April 2020 to Nov 2021.
We found that when we set a field as "Always one value selected" in a particular app, the Engine soon crashes.
We also tried using a backup saved before the upgrade, but the problem persist.
We didn't experience engine crashes with April 2020.
Is there a way to diagnose the problem?
Thank you very much,
Gianluca Ferri
Here is the CrashDump info, committed memory exceeds available memory. The app is not that big, it seems some kind of self reference, or something.
[ContextInfo]
Integrity=2
NbrActiveRequests=1
ReloadRequestId=-1
ExclusiveBlock=0
DocSxLock=0x0000000000000001
Username=xxxxxxxxxxxxxxxxxxxxx
SessionId=69eac605-f213-47c5-b4b1-d5a7fecda197
DocName=A87E58FA-670C-44EE-9959-7D281068C202
TenantId=default
[RequestInfo]
Integrity=3
OrderNumber=1053
TargetPtr=0x0000004C5A1BC118
TargetHandle=1
Exception=0
Started=1
Finished=0
Method=Doc::GetTablesAndKeys
SubObjectPath=
IsInValidate=0
[GlobalInfo]
Integrity=1
MemOverload=0.00
MemGCDepletion=0.00
MemLoLimit=24051487129
MemHiLimit=30923340595
MemCommitted=3192291328
MemReserved=4530139136
NbrCores=4
GcPageFileUsage=3150020608
OomErrors=0
Errno=34
StrErr=Result too large
ShutdownCalled=0
HasMTHigh=0
Hi, many reasons are cause of that.
1. sometimes it's happened because of conflict between similer standard formulas.
2. If synthetic key presented while table joined. Check data model, if any clear that
3. While developing stage some execution got stuck , that same app start using condition it would happen, so recreate that app once again.
4. Once struck or crash, without fail restart all services in server.
Hi. Good
I was faced same issue, those points listed by our implementation partner. Finally as you said we were landing inin memo issue. That also sudden memory growth which was not estimated growth. Finally we found unwanted log was not stopped from generation.
We finally found the problem: a numeric field contained a lot of NaNs and somehow this made the app unstable.
I believe this should be considered a bug, because it's very dangerous for the server.
Also, we did not have any problem with the same data in the previous version