Mobilink max upload

[eric_verhorstert]

SQL Anywhere 12.0.1.3797

Network protocol TCPIP

Speedtest internet (www.speedtest.net) location consolidated: 215 Mbps download, 430 Mbps upload

Size consolidate db 21Gb

Size consolidated log: 58Gb

Start Mobilink: mlsrv12.exe -v -c "DSN=<DSN> Client;UID=DBA;PWD=SQL" -cn 20 -w 10 -wm 50

Uploading data to our consolidated database fails when there is a lot of data. What we see in Mobilink monitor is that the "apply upload" keeps running and "never" stops. After 8 hours I shut down Mobilink to stop this connection. I also have the impression that I changed "something" that decreased the performance of Mobilink

For a upload that never finishes I unload the data to plain text files. The total number of bytes of these plain text files is 1.236.193.787 (1.15Gb).

* Is there a maximum number of bytes we can upload? Or, how can we find out why the "apply upload" keeps running.

* Is it possible to disconnect a remote id without shutting down Mobilink.

* Does the size of the log affect performance? Should we close down the database and rename the log?

* Concerning speed, i looked at a upload that finished without problems. I see: upload bytes 210282 apply upload: 1.194,322. Duration: 1.195,909

Anyone any ideas.

Thanks in advance,

Eric

[former_member188493 · Accepted solution]

The consolidated database log file does not affect MobiLink.

As far as the consolidated database server is concerned, the MobiLink server is just another client, and as such, MobiLink connections can be blocked... is that what's happening here?

upload bytes 210282 - that's "tiny" in the world of MobiLink synchronization.

Duration: 1.195,909 - if that's saying "over one thousand seconds", that's "forever" in the world of tiny MobiLink synchronizations.

So... you have dreadful performance. Please post more information, like the entire mlsrv12 -v+ -o diagnostic.txt file

Breck

[eric_verhorstert · Accepted solution]

Thanks Breck,

Performance suddenly decreased. About 2 records per second are processed.

Apply upload: 6.676,578 Upload bytes: 395069  Upload insert rows: 8218  Upload updated rows: 643

Today the message below appeared in the Mobilink console. Possibly because the leak of performance.

This happened only for 1 client.

W. 2014-04-10 15:41:34. <97> [10106] Unable to lock the remote ID <remote id>, will try again

W. 2014-04-10 15:41:35. <97> [10050] ODBC: [Sybase][ODBC Driver][SQL Anywhere]User 'DBA' has the row in 'ml_database' locked (ODBC State = 40001, Native error code = -210)

Is there an alternative to the -v+ because this will create a log file from several GB's. I also received the warning that the db consists of 5604 disk fragments. I should probably fix this first. Should we take the db down to fix this?

Computer info:

Processor: QEMU Virtual CPU version 2.00 GHz (4 processors)

Memory: 16Gb

System: Windows Server 2008 64 bts

Harddisk: 176 GB free

[former_member188493 · Accepted solution]

-vcefhkmnopstuU     logs everything interesting except data values

Did you check for blocking on the consolidated database?

SELECT * FROM sa_conn_info() WHERE blockedon <> 0;

There is a bottleneck, that's for sure, but more information is required. Foxhound might help as far as the consolidated database is concerned. The SQL Anywhere Monitor can show information about MobiLink.

How many syncs are happening at the same time?

[eric_verhorstert · Accepted solution]

Back in heaven. An index for one of the _del tables was dropped.

Added it and everything works fine again.