Difference between revisions of "User talk:Jzambor"

From Resco's Wiki
Jump to: navigation, search
(andrej)
 
(28 intermediate revisions by the same user not shown)
Line 1: Line 1:
  
==top==
+
<!--
  
<div style="display:inline-flex;width:275px;height:100px; background-color: #F7F7F7; margin: 0 10px 10px 0; ">
+
== Sample log files II ==
  <div style="width: 100px; height:100px; text-align: center; line-height:95px;">
 
    <div class="clickablecell" style="height:100px;">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
 
  </div>
 
  <div style="width: 175px; height:100px; line-height:95px;">
 
    <div class=" clickablecell " style="height:100px; font-size: 125%;">[https://academy.resco.net/ Resco Academy]
 
</div>
 
</div>
 
</div>
 
<div style="display:inline-flex;width:275px;height:100px; background-color: #F7F7F7; margin: 0 10px 10px 0; ">
 
  <div style="width: 100px; height:100px; text-align: center; line-height:95px;">
 
    <div class="clickablecell" style="height:100px;">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
 
  </div>
 
  <div style="width: 175px; height:100px; line-height:95px;">
 
    <div class=" clickablecell " style="height:100px; font-size: 125%;">[[Resco Cloud]]
 
</div>
 
</div>
 
</div><div style="display:inline-flex;width:275px;height:100px; background-color: #F7F7F7; margin: 0 10px 10px 0; ">
 
  <div style="width: 100px; height:100px; text-align: center; line-height:95px;">
 
    <div class="clickablecell" style="height:100px;">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
 
  </div>
 
  <div style="width: 175px; height:100px; line-height:95px;">
 
    <div class=" clickablecell " style="height:100px; font-size: 125%;">[[User interface components]]
 
</div></div>
 
</div>
 
<br>
 
  
<br>
+
<syntaxhighlight lang="xml">
==andrej==
+
                <SyncAnalyzer SkipEntitys='96' TotalTime='1076ms'/>
<div style="display:flex;flex-direction:row">
+
<ApiCalls Prepare='5' Uploader='1' DelPlugin='1' SyncAnalyzer='92' Downloader='1' NNDownloader='3' Total='104' />
    <div style="display: flex;margin: 0 1em 0 0;width: 275px;height: 100px">
+
</Summary>
        <div class="clickablecell">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
+
</syntaxhighlight>
        <div class=" clickablecell " style="height:100px; font-size: 125%;width: 100%;">[[Resco Cloud]]</div>
+
 
    </div>
+
Which statements are true?
    <div style="display: flex;margin: 0 1em 0 0;width: 275px;height: 100px">
+
- +If SyncAnalyzer was not used, we would save 1 API call. Lukas: pls explain, Sync analyzer used 92 Api calls, but there is 96 skipped entities, for which you would need to use 96 api calls, to get zero records, since they were not changed.
        <div class="clickablecell" style="height:100px;">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
+
 
        <div class=" clickablecell " style="height:100px; font-size: 125%;width: 100%;">[https://academy.resco.net/ Resco Academy]</div>
+
SyncAnalyzer concerns Downloader only.
    </div>
+
All excluded entities had to have 0 changes – otherwise they would not be excluded. SyncAnalyzer issued 1 call per entity. (FullSync counter need not be tested if IncSync returns small number of changes.) Same number of calls would be issued by the Downloader if SyncAnalyzer was disabled.
    <div style="display: flex;margin: 0 1em 0 0;width: 275px;height: 100px">
+
1 entity had some changes, but less than pagesize. (Because Downloader issued 1 call only.) This entity was tested by SyncAnalyzer, too (and FullSync counter was not queried because IncSync counter was small)– therefore 1 extra call.
        <div class="clickablecell" style="height:100px;">[[File:Wiki-icon-academy.png|link=lala|64px]]</div>
+
 
        <div class=" clickablecell " style="height:100px; font-size: 125%;width: 100%;">[[User interface component]] </div>
+
<syntaxhighlight lang="xml">
    </div>
+
<Settings App='13.1.2.0' Server='Crm2011.13/Federation' Threads='3' MaxExecMultiple='0' Foreground='false' Launch='Auto'/></syntaxhighlight>
</div>
+
 
 +
- +Upload will be slow when really needed ( 1 record/request.) Lukas: How can we assume this? Based on the slow download?
 +
 
 +
If we had 100 changes, MCRM would issue 100 web calls instead of one. Upload would be slower by 99 x web_latency, which might be anything between ~5 secs to ~50 secs.
 +
 
 +
<syntaxhighlight lang="xml"><Attachments Recv='6223' Skipped='1008' TotalTim='4236666ms' />
 +
</syntaxhighlight>
 +
 
 +
1008 attachments were not downloaded. These attachments cannot be displayed on the mobile device. Lukas: These 1008 are skipped, meaning, the max attachment size, or blob filter was used, so they are available in the app on demand, right?
 +
 
 +
Yes.
 +
 
 +
+The download performance is OK. If we subtract some time for server timeout, the performance will be a bit more than 1 ms/record, which is fine. Lukas: This is hard to assume, since we do not know how the SQL timeout is set. It can be set to 30 seconds or 5 minutes, and results will be different. We need to specify e.g. If we subtract 2 minutes for the server timeout from the download TotalTime…
 +
 
 +
Even if we assume 30 sec SqlServer timeout, we would have 191267 recs in 285 secs, which is performance 1.49 ms/rec. Still OK.
 +
 
 +
+Configuration setting ‘Max Sync Records’ is set to zero. (The setting does not make sense, perhaps used for testing only.) Lukas: How do you know that? I can only see that for entities, there was 0 records downloaded in IncSync, which can mean, that no changes were made on the server since the last sync.
 +
 
 +
The warning “Incomplete entities…” proves that the downloader was active and that it downloaded some records. All these records were refused because of MaxSyncCount settings. Because the downloader stored 0 records, that setting must have been 0.
 +
 
 +
 
 +
- +Extremally slow customization check. Should be investigated. Lukas: Isn’t it rather slow customization download? E.g. having a lot of Offline HTML files (like including media files)
 +
 
 +
No: First BkSync performs only check, no www/images download. Already this check took 37 secs.
 +
 
 +
- The use of SyncAnalyzer increased the number of API calls by 42. Lukas: The reason why this is not true is, that without SyncAnalyzer there will be more API calls, right?
 +
 
 +
SyncAnalyzer saves calls only in 1 situation: There are say millions of changed entity records, but only say thousands of entity records that match the sync filter.
 +
In this case we had small number of downloaded records. In such case SyncAnalyzer cannot save any web calls.
 +
 
 +
 
 +
<syntaxhighlight lang="xml">
 +
<FullSync Recv='17999' TotalTime='22047ms' Entitys='5407ms' ManyToMany='140ms'>
 +
<SharePointDownload REST='False' TotalTime='16032ms' GetListCollectionCalls='1x;1593ms' GetListCalls='1617x;13797ms' GetListItemsCalls='1x;109ms' IgnoredLists='2077' FileInfosSaved='1' />
 +
</syntaxhighlight>
 +
 
 +
This is a FullSync log, in which about ¼ of time was spent in entity download. Lukas: Please break down the calculation. Recv=17999 + SPDownload=16032 ~ time to sync. The recv part seems like ½ (this can be wrong, but why).  
 +
 
 +
(5407 + 140) / 22047
 +
17999 is the number of received records!
 +
 
 +
 
 +
 
 +
<syntaxhighlight lang="xml">
 +
</syntaxhighlight>
 +
 
 +
<syntaxhighlight lang="xml">
 +
</syntaxhighlight>

Latest revision as of 08:42, 20 November 2020