Skyp
-
Posts
76 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Posts posted by Skyp
-
-
On 1/22/2023 at 4:10 PM, Hayri ASLAN said:
Hello
We added this our backlog. We will solve this issue in the one of the next builds
We bypassed this error via TUnulayoutConfig.margin and Anchor in alignmentControl = Server. It's not convenient -(
Spent enough time to solve it somehow normally, tried a System.Classes change the sequence of reading Align and Margins, but this did not lead to anything.
When calling readState, there is an appeal to VCL.Controls (I could not go there from the call stack under Linux64) and after it all margins disappear (by default 3).
At the same time, I noticed a strange behavior in Margins.bottom for example (Unigroupbox), the bottom value does not disappear (by default 3), but is re-assigned to the Right, this can be seen in the Unisetbounds handler.
-
If VRRP works with hyperServer - I also think it could be a cool explanation in the FAQ on designing large systems, for example:
http://www.unigui.com/resources/online-documentation/developer-manual - Scalability(The issue of multiple DBMS connections is solved in my example PgBouncer)
Tomorrow morning I will try it 2 masters and one slave. Turn off-turn on . This may be an imitation of VRRP operation.. -
Hi.
We want to deploy our application on several datacenters, we are going to use the popular multiIP DNS routing(LoadBalancing) approach used for example in Google or large companies.
Ie, when a user sends an NS request, the nearest (one of the IP addresses) is returned to him via DNS, thereby the user will get to one of our datacenters (in Lithuania, Finland, Russia, etc.)..In each of the datacenters there are several hyper_server instances with VRRP(keepAlived) – a common IP, if one of the hyper_server instances refuses VIP service, the second instance will be responsible, while it will access the same nodes (slave) the same as the first one.
Because we plan to use the Rewrite load balancing mod and assign a unique white address to each working node (slave).
I.e. users come with NS: mydomain.com
Mydomain.com these are 3 IP addresses (different datacenters):
{
1. XXX.XXX.XXX.XXX – Finland(a)
2. XXX.XXX.XXX.XXX – Lithuania(b)
3. XXX.XXX.XXX.XXX – Russia ©
}
DataCenter #1(a) (for example Finland):
Common IP master – mydomain.com = XXX.XXX.XXX.XXX – Finland
{
Master#1 – keepAlived IP – Active
Master#2 – keepAlived IP – Reserve
Master#3 – keepAlived IP – Reserve
}
All master servers within the datacenter have the same configuration for working nodes (slave): [server-0], [server-1],[server-2],[server-3]..[server-n]
In Finland (slave), the nodes will be as follows:
Slave – subdomain.mydomain.com = XXX.XXX.XXX.XXX – Finland
{
Slave #1 - a0.mydomain.com
Slave #2 – a1.mydomain.com
Slave #3 – a2.mydomain.com
Slave #4 – a3.mydomain.com
... n
}(If one of the slave servers falls, it's not as scary as if the master falls and the automatic reserve does not rise.)
Then, of course, there is a VRRP for several PgBouncer (for a single pool of DBMS connections), followed by DBMS clusters with sharding and replication. (with auto-failower)All this architecture is not problematic except for one issue:
Will we be able to organize a hot backup via master hyper_server via VRRP? Did someone do it?How to automatically raise another master nearby (with the same IP – VIP)(he will immediately pick up all the nodes(slave VM)), does anyone have experience?
-
Good afternoon. The Mode-2 documentation doesn't say anything about remote Deployment on Linux in Mode2 (http://www.unigui.com/resources/online-documentation/developer-manual).
It doesn't seem to work in a Linux cluster.
I guess this error is related toI describe the sequence of steps (Step-By-Step):
Step 1 (i see fileinfo in Slave Server):
Step 3 (i see File info in Master server):
it's okey..
Step 4 I'm running services(first the slave, then the master):
And md5
Step 5 (i join in Master Server - /server):
it's ok
Step 6 (I'm uploading the image to master). Image i changed (recompile)
... Discard, purge, Active..
master is active - OK
Step 7 I'm moving to the Node state
it doesn't work =(
Step 8 - I go again and check the hash files on the master and slave
I see that the master has been updated and the slave has not.
Step 9 I don't give up and go to the slave - /server
oo...
Logs (slave):
when trying to log in, he writes Node is purged! (error 700)
What am I doing wrong? for me, this Hyper_server functionality is important
If you update the slave first and then the master, then the master does not see the slave - (the slave is in the Discarded status), only restarting the slave first and then the master helps
-
Hi, I wrote the instructions and within the framework of it I conducted testing on disabling the working node - receiving all sessions on the server (master) and enabling the working node.
The status of the node(VM) is updated to Active when enabled, but users whom the master should send to this node receive error 503. (Below I have described the Step-by-step sequence of steps)Step 1:
Step 2:
Step 3 (As if the node is out of order - fine):
Step 4 (I opened a large number of sessions and the master took them all on himself - that's fine):
Step 5 (I started the node😞
Step 6(the master's node status has changed to online - that's fine) :
Step 7 (I'm trying to open the interface (from Master)):
I specially reproduced this situation separately and attached Hyper_server logs. I ask for help how to exclude a situation when lifting the server (without restarting the Master) drops the work of a part of the cluster?
If, after enabling Node(Service/VM) (step 6), I restart and destroy all threads on the master, then the child node will work correctly and there will be no error 503 - but for this I will have to disconnect all users from the master:
Do you have any idea why this is happening?
Version: 1.90.0.1564
-
Linux does not catch the hyper_server version of the file as in the FAQ:
http://www.unigui.com/resources/online-documentation/developer-manual
I tried it in different ways - I press shift+F9 (Build) and it doesn't give a result, what to do?
-
Если кому-то актуально. Fix по UniDac Alerter (для Postgres):
В модуль PgClassesUni.pas
в часть uses добавляем модуль:
unit PgClassesUni; interface uses system.generics.collections,Находим класс TPgSQLNotificationsHandler(PgClassesUni.pas):
TPgSQLNotificationsHandler = class
private
FBaseConnection: TPgSQLConnection;
FConnection: TPgSQLConnection;
FEvents: array of TEventInfo;
FThread: TPgSQLNotificationsThread;
FMainTH: TThread; /// добавляем строчку
В этом-же классе процедуру:
DoOnNotificationПереносим на уровень public (декларативная часть)
Ниже описываем класс потока:
TMaintThread = class(TThread) strict private FQueue: TQueue<TPgSQLNotification>; FObj: TPgSQLNotificationsHandler; public procedure Execute; override; property Queue: TQueue<TPgSQLNotification> read FQueue write FQueue; property Obj: TPgSQLNotificationsHandler read FObj write FObj; end;
В тело Execute пишем:
{ TMaintThread } procedure TMaintThread.Execute; var li: integer; le: TPgSQLNotification; begin FreeOnTerminate := True; while not Terminated do begin sleep(100); if Assigned(FQueue) and Assigned(Obj) then begin for li := FQueue.Count - 1 downto 0 do begin le := FQueue.extract; Obj.DoOnNotification(le.PID, le.Name, le.Message); le.free; le:=nil; end; end; end; if Assigned(FQueue) then FreeAndNil(FQueue); Obj := nil; end;В Destroy добавляем :
if assigned(FMainTH) then FreeAndNil(FMainTH);Находим в этом же модуле процедуру:
procedure TPgSQLNotificationsHandler.ProcessNotification(const Name: string;
const PID: integer; const Message: string);
Заменяем:
Notif := TPgSQLNotification.Create; Notif.PID := PID; Notif.Name := Name; Notif.Message := Message; {$IFDEF MSWINDOWS} PostMessage(hUtilWindow, WM_NOTIFICATION, wParam(GCHandle), lParam(AllocGCHandle(Notif))); {$ELSE} try FLastNotification := Notif; FThread.Synchronize(DoNotification); finally FLastNotification := nil; Notif.Free; end; {$ENDIF}На вот это:
FLastNotification := TPgSQLNotification.Create; FLastNotification.PID := PID; FLastNotification.Name := name; FLastNotification.Message := Message; if not Assigned(FMainTH) then begin FMainTH := TMaintThread.Create(True); TMaintThread(FMainTH).Obj := Self; end; if not Assigned(TMaintThread(FMainTH).Queue) then TMaintThread(FMainTH).Queue := TQueue<TPgSQLNotification>.Create; TMaintThread(FMainTH).Queue.Enqueue(FLastNotification); FLastNotification := nil; if not FMainTH.Started then FMainTH.Start;
Наслаждаемся - все работает.
Причиной ошибки является вызов метода Synchronyze который фризит unigui сервер (ошибка возникает), этот метод нужен на случай если придёт большое кол-во сообщений (Synchronize - будет дожидаться основной отработки). В фиксе все сообщения отработают по очереди.
-
5 hours ago, Hayri ASLAN said:
Hello We added this our backlog. We will solve this issue in the one of the next builds
Right now I can't read Cookies when debugging on Centos (fix - check request access on save cookies) - I understand, in another week I came across the following problems - and solved it through crutches:
I'm already editing the source code of the framework, which is definitely wrong, can I get a quick crutch to solve this problem as well?
Can the problem be solved only on JS?(I don't have time to wait for a new build anyway)
-
There is a form on which there is a TvirtualTable (Unidac package) or Tfdmemtable (FireDAC package)(tried both).
uniDBCheckBox components bound to DataSource:
When changing the DataSource, I try to edit adjacent records in the same DataSet (code onAfterEdit):procedure TUFrameTreeRulesF.FFrmStructVTableAfterEdit(DataSet: TDataSet); procedure MainUpd; procedure UpdateRefresh(aIsEnabled, aIsNotifyInface, aIsNotifyEMail: Boolean; aParentStr: string); var lvq: TVirtualQuery; /// throught query (localSQL) lwherestr: string; const CDefUpdateParentStr : string = 'UPDATE source SET %s WHERE (idname=:pname OR idname LIKE :pname2) OR id=:pname;'; begin lwherestr := EmptyStr; lvq := TVirtualQuery.Create(nil); try lvq.SourceDataSets.Add(FFrmStruct.VTable, 'source'); if FIsEnabledChange then begin lwherestr := lwherestr + 'isEnabled = :pIsEnabled'; end; if FIsNotifyInfaceChange then begin lwherestr := ifthen(lwherestr.IsEmpty, EmptyStr, ',') + 'IsEnableNotifyInface =:pIsEnableNotifyInface'; end; if FIsNotifyEMailChange then begin lwherestr := ifthen(lwherestr.IsEmpty, EmptyStr, ',') + 'IsEnabledNotifyEmail=:pIsEnabledNotifyEmail'; end; if not lwherestr.IsEmpty then begin lvq.sql.Text := Format(CDefUpdateParentStr, [lwherestr]); lvq.ParamByName('pname').AsString := aParentStr; lvq.ParamByName('pname2').AsString := aParentStr + '%'; if FIsNotifyEMailChange then lvq.ParamByName('pIsEnabledNotifyEmail').AsBoolean := aIsNotifyEMail; if FIsNotifyInfaceChange then lvq.ParamByName('pIsEnableNotifyInface').AsBoolean := aIsNotifyInface; if FIsEnabledChange then lvq.ParamByName('pIsEnabled').AsBoolean := aIsEnabled; lvq.Prepare; try lvq.Execute; except on e: exception do lwherestr := e.message; /// check end; end; finally if Assigned(lvq) then lvq.Destroy; lvq := nil; end; end; var lubookmark: TBookmark; lparentid: string; begin FFrmStruct.VTable.AfterEdit := nil; lubookmark := FFrmStruct.VTable.Bookmark; FFrmStruct.UDbBuissnesObjectTree.BeginUpdate; FFrmStruct.VTable.DisableControls; FIsChangeSel := True; if not FFrmStruct.VTable.FieldByName('idname').AsString.IsEmpty then lparentid := FFrmStruct.VTable.FieldByName('idname').AsString else lparentid := FFrmStruct.VTable.FieldByName('id').AsString; FFrmStruct.VTable.edit; UpdateRefresh(UDBChBox.Checked, UDBCBNotifyInface.Checked, UDBCheckBoxNotifyEmail.Checked, lparentid); FFrmStruct.VTable.post; FFrmStruct.UDbBuissnesObjectTree.Refresh; FFrmStruct.VTable.EnableControls; FFrmStruct.UDbBuissnesObjectTree.EndUpdate; FFrmStruct.VTable.AfterEdit := FFrmStructVTableAfterEdit; if Assigned(lubookmark) then FFrmStruct.VTable.GotoBookmark(lubookmark); self.hidemask; end; begin self.ShowMask('Обновляю данные...'); FIsEnabledChange := not FFrmStruct.BeforeIsEnabledChange = UDBChBox.Checked; FIsNotifyInfaceChange := not FFrmStruct.BeforeisNotifyChange = UDBCBNotifyInface.Checked; FIsNotifyEMailChange := not FFrmStruct.BeforeIsNotifyEmail = UDBCheckBoxNotifyEmail.Checked; MainUpd; UBtnApplySetting.Enabled := True; end;After I get the error : "DataSet not in Edit or Insert mode" =(
At the same time, after the error occurred, the DataSet was updated (as expected, the code above worked (it works)). I have an error happening in the UniDbCheckBox.pas module:
I quickly corrected this via a "crutch" (hack):
Version: 1.90.0.1564
-
It is not possible to apply the triton theme,ubuntu, etc. not through the Theme Manager via Refresh, not in the designer. (Refresh doesn't work for some reason).
I fix it (Crutch(hack) :
I put the UniGuiConst.pas module next to the project and changed the Default_EXTHEME to the desired theme and it worked:
Version: 1.90.0.1564
PS:
ReCallLastTheme:=True; (but is not working in Linux (CentOS,RHEL) and works well in Windows).
-
On 1/19/2023 at 4:15 AM, Sherzod said:
Hello,
Edition? Still using trial edition?
Hi, do you have any mail angry Customer support or just a forum? the problem is very urgent-(
-
I added... there was a login
-
I added... there was a login
-
5 hours ago, Sherzod said:
Hello,
Edition? Still using trial edition?
Professional
-
Я установил значения в Design Time, и он отлично работает, если я строю проект под Windows :
,
.thumb.png.6bd1cb57e28271010131ef7972cafa38.png)
,
(StandAlone) В Windows (10) (это нормально):
.png.e70c1da91e2b3b5aab24e35f02fb6fd9.png)
,
(StandAlone) в Linux (CentOS 7) (это плохо = ():
.png.3856889c11d79da607ff0a003926d229.png)
,
Что я делаю неправильно? Разве это не работает под Linux? (alignwithmargins - маржа)
Вер: 1.90.0.1564
-
image on problem:
-
for trial use. For the task selection of tools for development.
-
Rad Studio Delphi xe8
UG 1.0.0 Build 1363 RC
Unidac 6.2.9
-
Hi.
Do not display values in BDGrid.
Processing fields in a query (q_for_users1) is as follows:// Query - good jobfor I := 0 to FieldCount-1 do
begin
s:='';
for j := 0 to pos('.',Field_list)-1 do
begin
s:=s+Field_list[j];
end;
for j := pos('.',Field_list)+1 to Length(Field_list) do
begin
s:=s+Field_list[j];
end;
with Q_for_users1 do
begin
Fields.Name:=Trim(s);
fields.DisplayLabel:=Field_list; // Field_list list of allowed characters
Fields.Visible:=true;
end;
end;Anddd insert a Grid:-------------------------------------------------------------------------------------------------------------------with UniMainModule do
begin
DATASOURCEUSER.Enabled := false;
DATASOURCEUSER.DataSet := Q_for_users1;
UniMainModule.DATASOURCEUSER.Enabled := true;
end;uniSession.AddJS((obj_out as TuniDBgrid).JSName+'.store.suspendEvents();');
(obj_out as TuniDBgrid).BeginUpdate;(obj_out as TuniDBgrid).DataSource := DATASOURCEUSER;
for i := 0 to Field_list.Count - 1 do
begin
if (obj_out is TuniDBgrid) = true then
begin
with (obj_out as TuniDBgrid) do
begin
(obj_out as TuniDBgrid).Columns.Add;
// s:= unimainmodule.Q_for_users1.fields.Name;
(obj_out as TuniDBgrid).Columns.Visible := true;
(obj_out as TuniDBgrid).Columns.FieldName :=Q_for_users1. fields.Name;
(obj_out as TuniDBgrid).Columns.title.Caption := Field_list;
(obj_out as TuniDBgrid).Columns.Width := 100;
end;
end;
end;
(obj_out as TuniDBgrid).endupdate;
uniSession.AddJS((obj_out as TuniDBgrid).JSName+'.store.resumeEvents();');
uniSession.AddJS((obj_out as TuniDBgrid).JSName+'.view.refresh();');
end;-------------------------------------------------------------------------------------
All runs without error but in the Grid cell (their number corresponds to the limit in the query) are empty. Who faced?
.png.76c8da494ba2b6db9823f173d8e87b2c.png)
,
Architectural question. VRRP and Master HyperServer. How?
in General
Posted
Does not work. I tried to deploy 2 Master servers, one turned off the second turned on with one slave – the cluster worked.
I turned off the 1st master, turned on the 2nd, and error 503 began to appear when routing to the node (slave).
node restart required (slave)....