Work Manager 6.2 on Windows Mobile 6.5 never completes synchronization

[Former Member]

Hi all.

Im unable to finish initial synchronization of Work Manager data on Windows Mobile device.

My landscape is as follows:

- Windows Mobile 6.5 device

- More than 512 MB of free space

- Work Manager 6.2 (downloaded from the Download center)

- SMP 3.0 SP08

The data in the devices and functional location tables is approx 100.000 records. The timeout on the server is set to 50 minutes. The internet connection is very good (both stable and fast). Certificate is successfully imported and used on the mobile device. Localisation is disabled (English language is used). Im installing Work Manager as Encrypted client data on the main partition of the device (although installing it on other partitions does not make any difference). Im not using camera nor scanner on device.

After I get connected and logged onto the server, the agentry starts to download the data (using Angel connection). The data keeps comming until a moment when, randomly, agentry displays either error 14 or error 11. The network did not fail at any point (I have monitored that), nor did the process last for more than 15 minutes before failing.

Reinstalling the app, cleaning the device, cleaning \data folder, clearing registry nor restarting it helps. At some point, the agentry continues to download the data and crashes in the middle of the process, without displaying any error. I also went through the SAP notes, and couldnt find anything that would resolve the issue.

Has anyone run into this kind of issue? Do you have any suggestions on how to proceed or debug it? Since there is nothing under "Windows Mobile 6.5" download link in SAP download center, I downloaded the only thing I could: application for Windows Mobile 6.0. Also, Windows Mobile 6.5 is on the official supported devices list (for some devices).

Thanks in advance!

Marko

[mark_pe · Accepted solution]

Marko,

When you said this: "After I get connected and logged onto the server, the agentry starts to download the data (using Angel connection). The data keeps comming until a moment when, randomly, agentry displays either error 14 or error 11."

Where is the point of failure, what is shown on the client? Are you getting a time out from downloading this functional location table or from other area?

If you open the Agentry log messages.log, there is an area that will show you the Failure "F" and timestamp. If you look at the messages.log what is the key during the point of failure or what was the steps before it failed? Was there too long of a transmit?

Error 14 means that you got cut or timed out.

You said you have 512 MB free space in the SD card or in memory?  Did you install the Agentry client in memory or in the SD card?  Suggestion is SD card.

Are you running the transmit WiFi, GPRS or LAN line?

Do you have any screen saver or power saver options on this device?

Did you try any other device aside from the device you are using?

How's your battery level on your Windows Mobile 6.5?

Let us know more information.

Best Regards,

Mark Pe
SAP Platinum Support Engineer

[Former Member · Accepted solution]

Hi Mark.

Thanks for your post. Based on your remarks, and in order to proceed, I did the following things:

- cleaned up the device (uninstalled app, deleted data folder, etc)

- cleaned server logs and restarted the server

- installed the SAP WM 6.2 application on the clean mobile device

- started Agentry client/app, entered the data and started sync.

The point of failure is hard to determine. Sometimes it occures while syncing functional locations, sometimes during ctUsers table, or at any other random point. The client shows either error 14 or 11. While doing the sync today, I was monitoring the network (I was connected on the same access point with another device, pinging the server where the SMP is installed) and the wasnt a single drop of packages.

Regarding a timeout, I never get any error which points to a timeout failure. At least not on the device.

Last three lines of the messages file on the SMP server are as follows:

S, 622, 21,  7, 11/13/2015 11:08:57,  67,21978350,  2771,      username, ANGEL: F959C0E8-B1C7-404C-94CC-B96D5F70D879

R, 622, 21,  7, 11/13/2015 11:09:03,  3,21978350,  2774,      username, ANGEL: F959C0E8-B1C7-404C-94CC-B96D5F70D879

F, 622, 21,  7, 11/13/2015 11:00:57,    ,21978350,  2774,      username, >unknown<

At the same time, last couple of lines on the SMP log file are:

2015 11 13 11:05:16#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processResults(JCO.Table)::begin |

2015 11 13 11:05:16#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processResults(JCO.Table)::rows returned:74964 |

2015 11 13 11:05:17#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processDeletedResults::begin |

2015 11 13 11:05:17#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processDeletedResults::rows returned:0 |

2015 11 13 11:05:17#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processDeletedResults(JCO.Table)::begin |

2015 11 13 11:05:17#+0100#INFO#System.out###Agentry SAPMM Worker Thread########processDeletedResults(JCO.Table)::rows returned:0 |

2015 11 13 11:05:20#+0100#INFO#System.out###Thread-2086########Session timer for user: USERNAME ending. Session locked: true Cancelled: false |

The last message, "Session timer for user: USERNAME ending", is always in the log when the agentry shows an error. Sometimes, this error jumps after only several seconds of syncing.

The same thing happens no matter where I install the application. For the todays testing I installed application in the \Application partition of the device. The device itself doesnt have SD card, but has built in partitions which are free to use. 3 out of 4 have plenty of free space.

Im connected via WiFi at the moment, but exactly the same happens when connected via GPRS. The same access point was used to sync Android and iOS devices, and produced no problems. Also, as I stated, Im monitoring the connection, as I connected another PC to the same access point, and the connection is solid.

Screen saver on the device is disabled and I keep checking the screen while the sync goes on, hoping to get an error that would indicate the problem. Its the same wether I connect the device to external power (via its craddle) or use it on its battery power. Battery level never dropped under 80%.

Regarding other devices:

- the same scenario, at the same time, works nicely on Android and iOS (iPhone and iPad) devices

- I have tried another Windows Mobile 6.5 device, with the same characteristics, and I got the same results

Where else could I look for errors? Can they be more descriptive? Are there any other tests I could try?

Thanks!

Marko

[bill_froelich · Accepted solution]

Marko,

What version of the Agentry client are you using?  Have you tried with the latest version?

--Bill

[mark_pe · Accepted solution]

Marko,

With respect to completing the transfer or transmit, your logs stated something:

Option 1:

Thread########processDeletedResults(JCO.Table)::rows

If the error normally happens on a delete steps this step normally checks the backend and compares it to your actual number of records on the device. The WinCE or Pocket PC Windows mobile 6.5 will utilize all its processing power to do the comparison.

If we compare this to your statement: "the same scenario, at the same time, works nicely on Android and iOS (iPhone and iPad) devices" those devices are much stronger for processing such task.  Eventually you may decide to use more powerful device than Windows Mobile 6.5.

Option 2:

Now to see if we can tweak or improve performance we need to analyze the delta of what is being deleted and compare this to your 100K records.

Our suggestion is to study how many records are in the table on how many is being deleted or there is another table in SMP that tracks all transmit ->  the SMP_CLUSTER_PROP_VALUE - if this table is seen to have lots of rows then your transmit will not finish. Please review: KBA # 2224520 - Performance Issue - SMP Query to dabase taking a couple seconds - large amount of old data in SMP_CLUSTER_PROP_VALUE.  This issue was fixed in SMP 3.0 SP09 PLX. You are using SMP 3.0 SP08.

Those are some of our suggestion.

Regards,

SAP Mobile Support Team

[Former Member · Accepted solution]

Hi Bill.

I have the latest version available on the SAP download web. Its SAP Work Manager 6.2.1.0 for Windows Mobile 6.0.

We have just finished SMP 3.0 upgrade to SP09 and Im going to do more tests on premises tomorrow. Hope to have some more information then. Will keep you posted.

Marko

[Former Member · Accepted solution]

You can find the latest WinCE client in the SMP SDK SP10.

[Former Member · Accepted solution]

Hi Stephen.

Thanks for suggestion. I did so, but, unfortunately, without any difference.

Marko

[Former Member · Accepted solution]

Hi Mark.

Thanks for options. Today I went on premise and tried several options. I downloaded latest SDK SP10, updated SMP 3.0 to SP09, and tested several client versions. I always cleaned up the device prior new installation (data, files, registry, start menu, ...). There was no difference from what happens on remote location.

Option 1:

- the CPU is A8 / 1 GHz. This should be more than enough for any kind of processing. Sometimes the error ocurs even on "small" tables, such as /SMERP/PM_CTFLOC_LSET_GET, which have less than 14k records.

Option 2:

- since you said that error was fixed in SP09, I upgraded from SMP 3.0 SP08 to SP09. Unfortunately, without any results. The step with the most data (75k) is /SMERP/PM_CTEQUIPMENT_GET. All other tables are smaller. The data is being transfered in 1k blocks.

Im doing some more tests today. Im trying to get the point of failure where it stops more often, or see anything that could give me more options to troubleshoot. Im monitoring both log and messages. Log gives me the latests executed BAPI, but the messages dont give anything except >unknown<.

Thanks!

Marko

[Former Member · Accepted solution]

I will point out that WinCE devices limit each application process to 32MB so if the client reaches that size the OS will not allow the process to grow any larger.

[Former Member · Accepted solution]

Hi Stephen.

How can I check that and what can I do to keep the client under 32 MB?

Marko

[mark_pe · Accepted solution]

Marko,

As stated above, you need to study "2224520 - Performance Issue - SMP Query to dabase taking a couple seconds - large amount of old data in SMP_CLUSTER_PROP_VALUE.  This issue was fixed in SMP 3.0 SP09 PLX. You are using SMP 3.0 SP08."

Note: Even though it stated as fix in higher version of the client, it is still recommended to do a query on how many lines (count) SMP_CLUSTER_PROP_VALUE has. This is known to slow down the transmit and can cause the failure of completing the transmit. Look at the example delete steps (you may need to state a date and time - older than what is needed). So if you have 500K rows (example only) in SMP_CLUSTER_PROP_VALUE - do you think the data listed in it (let us say you have data since 2014) is still needed.  What we have done with other customers is we queried this table and there were 2 million records of which only <50K may be needed. So after they delete the records they can complete their transmit or get rid of the error 11.

Hope this helps,

Mark Pe
SAP Platinum Support Engineer

[Former Member · Accepted solution]

Hi Mark.

Thanks for the reply. I will check the table. We are on Sybase, so Ill have to modify the stated query. But, I highly doubt that this is the issue, since its our test system, we dont have any data from 2014 nor any previous years. We have only 1 month of sample data in the system.

Marko

[chandra_ayyagari · Accepted solution]

Hi Marko,

Please try the following suggestions (process of elimination to find root cause). I still believe that your issue is related to response time like I previously stated. 

Step-1

Change the client global

a) Equipment.Worklist = N

b) FunctionalLocation.WorkList = N

Try to login and see if Transmit completes. If it does then please check the data volume of Classification Data - most likely this could be your issue because you have activated the client globals (/SMERP/PM_CTFLOC_LSET_GET is only executed if Global is True else standard BAPI is invoked).

Unfortunately the client wont show the exact CT it is loading when it becomes unresponsive. Sometimes triggering a parallel Transmit with different user (on different device iPhone) may help.

Step-2

Initially disable the complex tables related to Functional Location & Equipment (ENABLE_TABLE) parameter. Verify if you can complete Transmit (this should work given you disabled tables). Then reactivate the table and perform Transmit - your CTs should be downloaded on 2nd Transmit.

This could be a case of data volume you have to optimize SAP & delete unused indices on CT

Given it is working fine on iPhone and Android it is possible that we are hitting some sort of limit for Windows Mobile. If the 2-step approach of loading client works then you have to go with a pre-loaded client. 

On a side note it is not possible for end users to browse through or search for Equipment on a small screen unless you have barcoded the Equipment.

Thanks,

Chandra

[Former Member · Accepted solution]

Hi Chandra.

Thanks for your comments.

I have tried Step 1. I coppied GlobalsBase.ini to Globals.ini, in {sap}\MobilePlatform3\Server\configuration\com.sap.mobile.platform.server.agentry.application.SAPWM folder. I set the parameters as you suggested. I then cleared the device, installed everything back on, restarted the SMP server and ran the Agentry client on WinCE. Agentry connected and started downloading. It did download the functional locations I have in the system, despite the parameter [FUNCTIONALLOCATION]Worklist=N. It also downloaded lots of other tables and indexes before I got error 14.

Then, I resterted the device and restarted the client and started sync. Agentry client started to download everything again, as always. Functional locations got downloaded completely in the first attemp, and despite that, it started to download them again. Somewhere in the middle of the process (its 13k records) I got error 11.

Is there any other place where I could do any configuration? Am I checking the right configuration file?

Regarding step 2, I could find the enable_table parameter in the configuration. Should I just put it in "as is" or am I foreseeing something?

Thanks!

Update: I managed to do the initial sync, with disabling tables ctLocations and ctEquipment! Ill be investigating the issue further on, according to your guidelines, and let you know how it goes.

Marko

[Former Member · Accepted solution]

Hi Chandra.

I have finaly managed to do the initial sync and start the application. Step 2 from your suggestion was the only thing that worked. I set Enable_table parameter for Functional location and Equipment tables to 0. The sync then completed successfully. Then I switched it back to 1 to both tables, resynced and I have all the data on my mobile device.

It seems Ill have to do the two step sync for all Windows Mobile devices in this scenario.

Thanks to you and everyone for help!

Marko

[Former Member]

The scenario is now like this:

- SMP 3.0.9.3 SP09 Build: 20150916-1835

- Windows Mobile client is latest, taken from SDK SP10

I enabled debug logging in SMP for the WM application. This are the last lines from my SAPWM Worker Thread-# file:

2015/11/17 15:50:49.852:    
+ WorkFunction=0000000161FEF418 2015/11/17 15:50:49.852:       
+ User=SAPWM-user 2015/11/17 15:50:49.852:         
+ Application=SAPWM 2015/11/17 15:50:49.852:           
+ User=SAPWM-user 2015/11/17 15:50:49.852:             Originator exception: 2015/11/17 15:50:49.852:             
+ User=SAPWM-user 2015/11/17 15:50:49.852:               
+ ANGEL Connection=#.#.#.# 2015/11/17 15:50:49.852:                 Sending disconnect for vine 1129471106 2015/11/17 15:50:49.852:             Message failure case: 13 (Receiver Exception) 2015/11/17 15:50:49.852:             
+ User=SAPWM-user 2015/11/17 15:50:49.852:               Received LogItem Definition Request message 9 status changed to 'Failed'

Messages.log:

S, 206, 21,   9, 11/17/2015 15:50:49,  23, 20670,    29,       user, ANGEL: 142F6005-0A1D-42E8-8DF2-8EDF4F15DA91
R, 206, 21,   9, 11/17/2015 15:50:49,   0, 20670,    31,       user, ANGEL: 142F6005-0A1D-42E8-8DF2-8EDF4F15DA91
S, 206, 21,   9, 11/17/2015 15:50:49,  23, 31972,    31,       user, ANGEL: 142F6005-0A1D-42E8-8DF2-8EDF4F15DA91
R, 206, 21,   9, 11/17/2015 15:50:49,   3, 31972,    34,       user, ANGEL: 142F6005-0A1D-42E8-8DF2-8EDF4F15DA91
F, 206, 21,   9, 11/17/2015 15:50:47,    , 31972,    34,       user, >unknown<

SMP log file:

2015 11 17 15:50:49#+0100#DEBUG#com.sap.mobile.platform.server.agentry.core###Agentry SAPWM Worker Thread########Sending 11314 bytes of data |
2015 11 17 15:50:49#+0100#DEBUG#com.sap.mobile.platform.server.agentry.core###http-bio-8081-exec-2########Receiving 15 bytes via ByteBuffer backing array |
2015 11 17 15:50:49#+0100#DEBUG#com.sap.mobile.platform.server.agentry.core###http-bio-8081-exec-2########Done receiving data |
2015 11 17 15:50:49#+0100#DEBUG#com.sap.mobile.platform.server.agentry.core###Agentry SAPWM Worker Thread########Sending 12 bytes of data |
2015 11 17 15:50:49#+0100#DEBUG#com.sap.mobile.platform.server.agentry.console###Agentry SAPWM Worker Thread########Received LogItem Definition Request(206), 21, 9, user, 31972/34, 3:50:47PM/3:50:47PM |

Hope this gives someone an idea. The process now stops at "Replacing definition for Register Periodic Add" section and no errors except the stated are shown. The Agentry mobile application gives Error 11. After that, it refuses to connect (issuing error 14) untill I exit and start the application again.

Thanks in advance!

Marko

[LeonardoGomez]

Hi,

I'm having this issue on a new customer. It looks the same. SMP, Work Manager and SDK versions are the same. I'm testing with WPF client and getting error 11 in different points, most commonly after finishing Get Workorders, but sometimes it gets to Get Notifications. Server logs looks very similar to those you just posted.

I checked with debugging that only 1 workorder and no notifications are being found.

Please, let me know if there are any updates for this thread.

Thanks!

Leonardo.

[Former Member]

Hi Leonardo.

I still havent been able to get the initial sync done and the behaviour is similar to yours. I have tried almost everything from the list above, I have an open message to SAP support where I have also tried almost every option, but still without success.

I will be working more on this issue and share any conclusions to which I might come.

Marko

[LeonardoGomez]

Thanks Marko. I'll be checking for updates on this thread.

If you don't mind, you can also share your SAP support ticket with me as I can get into the system there.

Kind regards,

Leonardo.

[Former Member]

Hi Leonardo.

Chandra suggested to set ENABLE_TABLE parameter from 1 to 0, for tables Functional Locations and Equipment. This helped me get the initial sync and I could start the application for the first time. Then, I set the parameters back to 1 and did another sync, and this got me full sync of the data to the mobile device.

Hope this will help you in your case.

Marko

[LeonardoGomez]

Hello Marko,

Thanks for the info. Unfortunately in my case is a little bit different:

- I am getting only communication error 11, not 14

- The moment of accurance is not random. It always appear after the workorders are downloaded to the client. If there isn't any workorders for the user it appears after the notifications.

- If I sync right after the error, it resumes and the client opens. But if I remove the workorders and sync again it will get the error as in the beginning.

I'll follow this with OSS.

Thanks!

Leonardo.

[bill_froelich]

Marko,

What does the connection look like between the device and the SMP server?  Are you on the same local network or are you connecting through the Internet, VPN, firewalls, proxies etc?

Typically error 14 during a transmit points to a network timeout where something between the client and SMP server is closing the websockets connection while the transmit is still in progress.  This is especially true when connecting through firewalls and proxies.

If you are external, can you get the device on the same network as the SMP server and try the synchronization that way to see if it will complete?

--Bill

[chandra_ayyagari]

Please check the Timeout setting on SMP Server and compare to the time it takes to execute the BAPI in backend. If you are BAPI takes long then you can get disconnected.

We have loaded over 100k records to Complex Tables but in general for Windows PPC devices it is better to restrict the data to 50 or 60K.

thanks

Chandra   

[Former Member]

Hi Bill.

Thanks for you comments.

Im connecting through internet. The link is very stable and fast (30 MBit, fiber optic). We do have a firewall set up, no proxies and no VPN. I can try syncing from the SMP location, but I highly doubt this is the issue, as my other devices (Android and iOS) connect and sync without any issues or errors.

Sometimes, after I restart the sync process, the error occurs after less than a minute of data transfer. I checked the numbers on the server side, and I set them, for test purposes, very high (timeouts in hours).

Marko

[Former Member]

Hi Chandra.

Thanks for thoughts. I checked the timeout numbers and they are very high. Im pretty sure its not the issue, since the sync goes nicely on Android and iOS devices.

Marko

[bill_froelich]

My only other suggestion would be to test with the latest client (SDK SP10 and the latest patch release) to see if that makes any difference.  If not you will probably need to open a ticket with support.

--Bill

[Former Member]

H Bill.

I did this today. I downloaded latest client, as suggested (SDK SP10) and it makes absolutely no difference. I also opened a ticket with support, but without any info so far.

Thanks for the ideas!

Marko