vmware - BizTalk connectivity issue to SQL during VM snapshot -

vmware - BizTalk connectivity issue to SQL during VM snapshot -

we have 1 vm biztalk , separate vm sql backend. using veeam backups kicks off snapshot of vm. when snapshot beingness finalized on sql vm, biztalk services on application server fail. restart automatically requires manual intervention start services. error below logged on biztalk server.

is there timeout setting or config changes allow biztalk services remain during snapshot process?

an error occurred requires biztalk service terminate. mutual causes following: 1) unexpected out of memory error. or 2) inability connect or loss of connectivity 1 of biztalk databases. service shutdown , auto-restart in 1 minute. if problematic database remains unavailable, cycle repeat.

error message: [dbnetlib][connectionread (recv()).]general network error. check network documentation. error source:

biztalk host name: biztalkserverapplication windows service name: btssvc$biztalkserverapplication

we experienced same situation , error both biztalk 2009 , biztalk 2013, each set 2 app servers , 1 sql db server.

when our vmware final step of snapshot backup on application servers, freezes application server 10 seconds, preventing receiving packets. on sql server 2008 , 2012, default send out keep-alive packets clients every 30 seconds (30,000 ms). if sql server fails receive response app server, send out 5 retries (default setting) of keep-alive request 1 sec (1,000 ms) apart. if sql still not receive response back, terminate connection, cause biztalk hosts on app server reset, , in our case, when our german-made erp scheme sends edi documents on biztalk during reset period, transmission fail.

we trapped issue running netmon on db , app servers, waiting next error message. upon inspection, see 5 sql keep-alive packets beingness sent app servers 1 sec apart, , @ same time there no packets @ received on application server. @ first guess, 1 might think "just dropped network packets", case. made correlation timing of vm snapshots, , confirm each time snapshot finishes each day, app servers freeze.

as short-to-mid-term workaround, raised number of retries sql attempts before declaring connection dead, (5 default), adding registry value tcpmaxdataretransmissions , setting 30 (thus 30 seconds before sql declares client unresponsive). has masked problem us, , utilize @ own discretion.

we looking @ agent-based version of vm snapshot, may alleviate status of freezing server.

sql vmware biztalk