nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

UAPI AddToByIDs strange behavior

dluo

Hopefully I am in the correct forum to post this.

We have an issue with the call to UAPI.AddToByIDs. The function is called within an event script with admin session created in the code. When ever the call is used to add the current step user (or previous step user if it's called in an automatic step) into a group, the user will experience a hanging ~ 2 minutes before the web page return to the assignment page.

The affected workflow and code are upgraded from LL9.7.1 to CS10.5. It used to work without problem. To temporarily solve the problem, we added an automatic step after the user step and before the event script call. This step does nothing and is assigned to an agent user. It seems work around the issue. But we'd like to know are there any changes in the implementation of UAPI.AddToByIDs between different versions. What should be the appropriate scenarios to use it in CS10.5?

Thanks.

Dong

Developer

Stellar Services.

Find more posts tagged with

Comments

eLink User

On a related development thread one of the ex senior OT developers indicated something like this which matches your description.

Assume you had the event script made in builder 9.71 and when you moved into 10.5 you probably got the module shell and

imported the legacy module and called it working.However if you hadn't completely recompiled every script one by one in 10.5

performance degradation could occur. I am talking about a case where you would have just called your eventscript_9_7_1 as eventscript_10_5_0 changed the kernel and started using it.What I think this developer was hinting is you would re-evaluate the 971 oll and issue new oll code with CSIDE.

Sorry I am not the developer who is behind this info, and when I ported my 971 eventscripts I recompiled them in 10.5.

Secondly use of the administrative context is very tricky in newer versions.For e.g gimmeaprgctx will throw errors if there is no database connections and oscript will issue a warning.So revisit that code or put echos to see if you are popping valid sessions off the stack.you should in my mind be able to put ticks out in your script enter and exit and see if you can profile them in 9.7.1 and 10 or later.

.....appu appnair appunair

Well, if I called the wrongnumber, why did you answer the phone?
James Thurber, New Yorker cartoon caption, June 5, 1937

On Thu, May 26, 2016 at 9:22 AM, eLink Entry: Content Server Development Forum <development@elinkkc.opentext.com> wrote:

UAPI AddToByIDs strange behavior Posted byLuo, DongOn 05/26/2016 10:15 AM

Hopefully I am in the correct forum to post this.We have an issue with the call to UAPI.AddToByIDs. The function is called within an event script with admin session created in the code. When ever the call is used to add the current step user (or previous step user if it's called in an automatic step) into a group, the user will experience a hanging ~ 2 minutes before the web page return to the assignment page.The affected workflow and code are upgraded from LL9.7.1 to CS10.5. It used to work without problem. To temporarily solve the problem, we added an automatic step after the user step and before the event script call. This step does nothing and is assigned to an agent user. It seems work around the issue. But we'd like to know are there any changes in the implementation of UAPI.AddToByIDs between different versions. What should be the appropriate scenarios to use it in CS10.5?Thanks.DongDeveloperStellar Services.
[To post a comment, use the normal reply function]
Forum: Content Server Development Forum
Content Server: Knowledge Center

dluo

Thanks for the reply. But our issue is not exactly match the case.

The custom module is already recompiled for CS10.5 in eclipse. To make it clear, we did it like this:

1. change the kernel dependency in ini file, install it in CS10.5

2. import the code in eclipse, make necessary changes, so that the code can be compiled without any error messages.

After that, I assume the event script is already compiled for CS 10.5.

The performance issue only happens for two particular users. When logged in as the affected user, we can repeat the problem by:

1. create a test group

2. add current user into the new group via web interface

Here, no custom code is invovled, but the problem is the same.

The affected users are already added into a large number of groups. These groups are roles of binders. Other users that are not added to large number of groups don't have this problem.

It also takes ~ 2 minutes for the affected users to log in to CS10.5 whenever their session are timed out.

Both performance problems did not occur in LL9.7.1 system.

Dong

eLink User

unfortunately this is for you to take up directly with OT through logs(connect and thread logs) The product management responsible for TW will work on your

problems. The project kind of code (role based) has in fact many times been re-worked for better performance. TW objects inherit a lot of role based code so your

input and case will help make them do better code eventually.

If you we re going to fight it on your own I would basically look at the connect statements and show it to a DBA unless you are one yourself who could tell if better queries could be constructed.

.....appu appnair appunair

Well, if I called the wrongnumber, why did you answer the phone?
James Thurber, New Yorker cartoon caption, June 5, 1937

On Tue, May 31, 2016 at 1:46 PM, eLink Entry: Content Server Development Forum <development@elinkkc.opentext.com> wrote:

Re UAPI AddToByIDs strange behavior Posted byLuo, DongOn 05/31/2016 02:44 PM

Thanks for the reply. But our issue is not exactly match the case.The custom module is already recompiled for CS10.5 in eclipse. To make it clear, we did it like this:1. change the kernel dependency in ini file, install it in CS10.52. import the code in eclipse, make necessary changes, so that the code can be compiled without any error messages.After that, I assume the event script is already compiled for CS 10.5.The performance issue only happens for two particular users. When logged in as the affected user, we can repeat the problem by:1. create a test group2. add current user into the new group via web interfaceHere, no custom code is invovled, but the problem is the same.The affected users are already added into a large number of groups. These groups are roles of binders. Other users that are not added to large number of groups don't have this problem. It also takes ~ 2 minutes for the affected users to log in to CS10.5 whenever their session are timed out.Both performance problems did not occur in LL9.7.1 system.Dong
[To post a comment, use the normal reply function]
Topic: UAPI AddToByIDs strange behavior
Forum: Content Server Development Forum
Content Server: Knowledge Center

dluo

Yes, we opened a ticket with OpenText. In the immediate reply, your answer in this post is referenced. That's why I come back here to look for more help. Can you kindly give the instruction where to look at the connect statement?

Thanks.

Dong

Hans Stoop

I don’t know which update you are on, but you should also go to update 201409. Considering that you are currently doing an upgrade to 10.5, I assume you installed a later update.

In how many roles are these users?

Do they really have to be in roles or shouldn’t they not just be added via a normal group to the ACL.

I have seen customers that have already prefilled the roles on the template. That is very bad practice. These should be replaced by normal groups.

After a binders is closed you should consider removing all roles and replace them by groups.

Hans

From: eLink Entry: Content Server Development Forum [mailto:development@elinkkc.opentext.com]
Sent: Dienstag, 31. Mai 2016 20:46
To: eLink Recipient
Subject: Re UAPI AddToByIDs strange behavior

Re UAPI AddToByIDs strange behavior

Posted by Luo, Dong On 05/31/2016 02:44 PM

Thanks for the reply. But our issue is not exactly match the case.

The custom module is already recompiled for CS10.5 in eclipse. To make it clear, we did it like this:

1. change the kernel dependency in ini file, install it in CS10.5

2. import the code in eclipse, make necessary changes, so that the code can be compiled without any error messages.

After that, I assume the event script is already compiled for CS 10.5.

The performance issue only happens for two particular users. When logged in as the affected user, we can repeat the problem by:

1. create a test group

2. add current user into the new group via web interface

Here, no custom code is invovled, but the problem is the same.

The affected users are already added into a large number of groups. These groups are roles of binders. Other users that are not added to large number of groups don't have this problem.

It also takes ~ 2 minutes for the affected users to log in to CS10.5 whenever their session are timed out.

Both performance problems did not occur in LL9.7.1 system.

Dong

[To post a comment, use the normal reply function]
Topic:	UAPI AddToByIDs strange behavior
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

eLink User

Hans stoop of OT a very respected ,member of OT knows the product has given you some suggestions.I hope you adhere to what Hans says.

For academic interest this is how traditional logging works.

in the opentext.ini one goes and

changes

DEBUG=2

wantLogs=true

As programmers we always re-start or do it from ot.ini because we can then see the thread loading info.Note you can use the

GUI to do this as well ?func=admin.sysvars

the above keys in ot.ini will show up in GUI as Detailed Thread logging+SQL Logging

This combination enhances thread logging and connect logging

In newer livelink servers you have modern log4xxx logging that can work without re-starts et al.

Kyle Swidorwich has an article published that will explain the newer(java style of logging) styles.

if your server has 5 threads then thread1.out+connect1.out is how you would know what is doing what.

.....appu appnair appunair

Well, if I called the wrongnumber, why did you answer the phone?
James Thurber, New Yorker cartoon caption, June 5, 1937

On Tue, May 31, 2016 at 2:23 PM, eLink Entry: Content Server Development Forum <development@elinkkc.opentext.com> wrote:

Re Re UAPI AddToByIDs strange behavior Posted byLuo, DongOn 05/31/2016 03:22 PM

Yes, we opened a ticket with OpenText. In the immediate reply, your answer in this post is referenced. That's why I come back here to look for more help. Can you kindly give the instruction where to look at the connect statement?Thanks.Dong
[To post a comment, use the normal reply function]Topic:UAPI AddToByIDs strange behaviorForum:Content Server Development ForumContent Server:Knowledge Center

Hans Stoop

<![if !supportLists]>Ø <![endif]>we can then see the thread loading info

With one of the latest updates the tread loading is always logged in a separate log file.

Maybe only when debugging is not set already.

Hans

From: eLink Entry: Content Server Development Forum [mailto:development@elinkkc.opentext.com]
Sent: Dienstag, 31. Mai 2016 21:51
To: eLink Recipient
Subject: Re Re Re UAPI AddToByIDs strange behavior

Re Re Re UAPI AddToByIDs strange behavior

Posted by eLink On 05/31/2016 03:51 PM

Hans stoop of OT a very respected ,member of OT knows the product has given you some suggestions.I hope you adhere to what Hans says.

For academic interest this is how traditional logging works.

in the opentext.ini one goes and

changes

DEBUG=2

wantLogs=true

As programmers we always re-start or do it from ot.ini because we can then see the thread loading info.Note you can use the

GUI to do this as well ?func=admin.sysvars

the above keys in ot.ini will show up in GUI as Detailed Thread logging+SQL Logging

This combination enhances thread logging and connect logging

In newer livelink servers you have modern log4xxx logging that can work without re-starts et al.

Kyle Swidorwich has an article published that will explain the newer(java style of logging) styles.

if your server has 5 threads then thread1.out+connect1.out is how you would know what is doing what.

.....appu appnair appunair

Well, if I called the wrong number, why did you answer the phone?
James Thurber, New Yorker cartoon caption, June 5, 1937

On Tue, May 31, 2016 at 2:23 PM, eLink Entry: Content Server Development Forum <development@elinkkc.opentext.com> wrote:

Re Re UAPI AddToByIDs strange behavior

Posted by Luo, Dong On 05/31/2016 03:22 PM

Thanks.

Dong

[To post a comment, use the normal reply function]
Topic:	UAPI AddToByIDs strange behavior
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

[To post a comment, use the normal reply function]
Topic:	UAPI AddToByIDs strange behavior
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

dluo

Recently the content server service is restarted. Now the affected user is experiencing slow performance even navigating in content server. Other users don't have this problem.

I do think the suggestion here will be the solution to our performance problem. Are there anything to pay attention to when replace binder role with normal group?

dluo

BTW, we did collect the logs with debug options turned on.

However, I did not see any output during the hanging when it happened. The whole system including the console output just stop response for ~ 2 minutes. As soon as the system is back, the log is back too.

John Simon

In the connect logs you should find statements like these:

06/20/2016 07:59:35 INFO [3246544816] 0000006496: KSql::Execute(...) --> 'SUCCESS',1 record,[sec: 0 msec: 1]

They show you how long the SQL (immediately preceding) took to run...this particular one took 1 millisecond. You shold be able to search thru the connect logs and find any SQL taking a long time to run. I find it always helps pinpoint these kinds of problems.

Cheers...

dluo

Thanks, this is what I found in the connect log related to the issue:

06/03/2016 12:35:32 INFO [218011833858] 0000299276: KSql::Execute(...) --> 'SUCCESS',1 record,[sec: 115 msec: 92]
06/03/2016 12:35:32 INFO [218011833917] 0000299277: KLogin::PopulateKUAFRightsListNew(218858) --> 'SUCCESS'

This long running function, 'PopulateKUAFRightsListNew', is only executed when affected user is adding himself into a group. When he is adding another user or another user is adding him, this function is not triggered.

Dong

dluo

The full log related should be:

06/03/2016 12:33:37 INFO [217896741133] 0000298937: KSqlCursor::Open(4000,'DECLARE @ChildID BIGINT; SET @ChildID=:A1; WITH KUAFChildrenCTE(RLID, ChildID, RLRightID, RLProxyType, ChildPath) AS ( SELECT @ChildID AS RLID, ChildID, ID AS RLRightID, 0 AS RLProxyType, CAST( '|' + CAST(ChildID AS VARCHAR) + '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren WHERE ChildID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, ID AS ChildID, ID AS RLRightID, ProxyType AS RLProxyType, CAST('|' + CAST( @ChildID AS VARCHAR)+ '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFProxy WHERE ProxyID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, E.ChildID, E.ID, M.RLProxyType, CAST( M.ChildPath + CAST(E.ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren AS E join KUAFChildrenCTE AS M ON E.ChildID = M.RLRightID WHERE NOT M.ChildPath LIKE ('%|'+CAST(E.ID AS VARCHAR)+'|%')) INSERT INTO KUAFRightsListNew SELECT * FROM ( SELECT RLID, RLRightID, RLProxyType FROM KUAFChildrenCTE UNION ALL SELECT @ChildID, @ChildID, 0 UNION ALL SELECT @ChildID, -1, 0 UNION ALL SELECT @ChildID, ProxyID, ProxyType FROM KUAFProxy WHERE ProxyID=@ChildID ) AS SOURCE EXCEPT SELECT * FROM KUAFRightsListNew WHERE RLID = @ChildID OPTION (MAXRECURSION 0);WITH KUAFChildrenCTE(RLID, ChildID, RLRightID, RLProxyType, ChildPath) AS ( SELECT @ChildID AS RLID, ChildID, ID AS RLRightID, 0 AS RLProxyType, CAST( '|' + CAST(ChildID AS VARCHAR) + '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren WHERE ChildID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, ID AS ChildID, ID AS RLRightID, ProxyType AS RLProxyType, CAST('|' + CAST( @ChildID AS VARCHAR)+ '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFProxy WHERE ProxyID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, E.ChildID, E.ID, M.RLProxyType, CAST( M.ChildPath + CAST(E.ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren AS E join KUAFChildrenCTE AS M ON E.ChildID = M.RLRightID WHERE NOT M.ChildPath LIKE ('%|'+CAST(E.ID AS VARCHAR)+'|%')) DELETE FROM KUAFRightsListNew FROM ( SELECT * FROM KUAFRightsListNew WHERE RLID = @ChildID EXCEPT SELECT * FROM ( SELECT RLID, RLRightID, RLProxyType FROM KUAFChildrenCTE UNION ALL SELECT @ChildID, @ChildID, 0 UNION ALL SELECT @ChildID, -1, 0 UNION ALL SELECT @ChildID, ProxyID, ProxyType FROM KUAFProxy WHERE ProxyID=@ChildID ) AS SOURCE ) AS todel WHERE todel.RLID = KUAFRightsListNew.RLID and todel.RLRightID = KUAFRightsListNew.RLRightID and todel.RLProxyType = KUAFRightsListNew.RLProxyType OPTION (MAXRECURSION 0);') -->'SUCCESS'
06/03/2016 12:33:37 INFO [217896741420] 0000298938: KSqlCursor::Prepare() --> 'SUCCESS'
06/03/2016 12:33:37 INFO [217896741581] 0000298939: KSqlCursor::Bind(1,{218858}) --> 'SUCCESS'
06/03/2016 12:35:32 INFO ..
06/03/2016 12:35:32 INFO [218011833117] 0000299274: KSqlCursor::Execute() --> 'SUCCESS'
06/03/2016 12:35:32 INFO [218011833764] 0000299275: KSqlCursor::Close() --> 'SUCCESS'
06/03/2016 12:35:32 INFO [218011833858] 0000299276: KSql::Execute(...) --> 'SUCCESS',1 record,[sec: 115 msec: 92]
06/03/2016 12:35:32 INFO [218011833917] 0000299277: KLogin::PopulateKUAFRightsListNew(218858) --> 'SUCCESS'

dluo

@Hans Stoop,

Can you elabrate more on how to achive this?

"After a binders is closed you should consider removing all roles and replace them by groups."

The answer I received from OT support is:

"After The roles themselves (Coordinator, members and Guests) can't be removed. You can only add groups to them."

eLink User

In some situations we can run the offending query against a SQLtool. In Oracle you will hear such terms as AWR and traces.The simplest would be to run this in a SQLtool and Oracle and by extension SQLserver also should have profiling tools.In many cases you will with the help of a good DBA (Since John Simon is not working for you

) be able to speed that up.In your particular case We regularly try to involve DBA's and they will many times identify things to do(some examples get statistics,add indexes,plan execution,text indexes etc etc) Hans Stoop has said that OT is aware of the inefficiencies and asked you to consider doing what they think you should be doing .In case you could not get perhaps new code or changed processes one could consider optimizing with one's own knowledge.Sometimes one could even write a better statement and put it where(patch,override etc these are advanced oScript topics) the slowness is.If I were you on another Dev database and a LLserver and I would put the version that Hans Stoop said you should get into and see if that is faster.

Anyways the 115 secs is the 2 minute you correctly identified

Awesome isn't it the debugging help OT provides ...

.....appu appnair appunair

Well, if I called the wrongnumber, why did you answer the phone?
James Thurber, New Yorker cartoon caption, June 5, 1937

On Mon, Jun 20, 2016 at 10:27 AM, eLink Entry: Content Server Development Forum <development@elinkkc.opentext.com> wrote:

RE Re UAPI AddToByIDs strange behavior Posted byLuo, DongOn 06/20/2016 11:19 AM

The full log related should be:06/03/2016 12:33:37 INFO [217896741133] 0000298937: KSqlCursor::Open(4000,'DECLARE @ChildID BIGINT; SET @ChildID=:A1; WITH KUAFChildrenCTE(RLID, ChildID, RLRightID, RLProxyType, ChildPath) AS ( SELECT @ChildID AS RLID, ChildID, ID AS RLRightID, 0 AS RLProxyType, CAST( '|' + CAST(ChildID AS VARCHAR) + '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren WHERE ChildID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, ID AS ChildID, ID AS RLRightID, ProxyType AS RLProxyType, CAST('|' + CAST( @ChildID AS VARCHAR)+ '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFProxy WHERE ProxyID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, E.ChildID, E.ID, M.RLProxyType, CAST( M.ChildPath + CAST(E.ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren AS E join KUAFChildrenCTE AS M ON E.ChildID = M.RLRightID WHERE NOT M.ChildPath LIKE ('%|'+CAST(E.ID AS VARCHAR)+'|%')) INSERT INTO KUAFRightsListNew SELECT * FROM ( SELECT RLID, RLRightID, RLProxyType FROM KUAFChildrenCTE UNION ALL SELECT @ChildID, @ChildID, 0 UNION ALL SELECT @ChildID, -1, 0 UNION ALL SELECT @ChildID, ProxyID, ProxyType FROM KUAFProxy WHERE ProxyID=@ChildID ) AS SOURCE EXCEPT SELECT * FROM KUAFRightsListNew WHERE RLID = @ChildID OPTION (MAXRECURSION 0);WITH KUAFChildrenCTE(RLID, ChildID, RLRightID, RLProxyType, ChildPath) AS ( SELECT @ChildID AS RLID, ChildID, ID AS RLRightID, 0 AS RLProxyType, CAST( '|' + CAST(ChildID AS VARCHAR) + '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren WHERE ChildID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, ID AS ChildID, ID AS RLRightID, ProxyType AS RLProxyType, CAST('|' + CAST( @ChildID AS VARCHAR)+ '|' + CAST(ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFProxy WHERE ProxyID in (@ChildID) UNION ALL SELECT @ChildID AS RLID, E.ChildID, E.ID, M.RLProxyType, CAST( M.ChildPath + CAST(E.ID AS VARCHAR) + '|' AS VARCHAR(MAX)) FROM KUAFChildren AS E join KUAFChildrenCTE AS M ON E.ChildID = M.RLRightID WHERE NOT M.ChildPath LIKE ('%|'+CAST(E.ID AS VARCHAR)+'|%')) DELETE FROM KUAFRightsListNew FROM ( SELECT * FROM KUAFRightsListNew WHERE RLID = @ChildID EXCEPT SELECT * FROM ( SELECT RLID, RLRightID, RLProxyType FROM KUAFChildrenCTE UNION ALL SELECT @ChildID, @ChildID, 0 UNION ALL SELECT @ChildID, -1, 0 UNION ALL SELECT @ChildID, ProxyID, ProxyType FROM KUAFProxy WHERE ProxyID=@ChildID ) AS SOURCE ) AS todel WHERE todel.RLID = KUAFRightsListNew.RLID and todel.RLRightID = KUAFRightsListNew.RLRightID and todel.RLProxyType = KUAFRightsListNew.RLProxyType OPTION (MAXRECURSION 0);') -->'SUCCESS'
06/03/2016 12:33:37 INFO [217896741420] 0000298938: KSqlCursor::Prepare() --> 'SUCCESS'
06/03/2016 12:33:37 INFO [217896741581] 0000298939: KSqlCursor::Bind(1,{218858}) --> 'SUCCESS'
06/03/2016 12:35:32 INFO ..
06/03/2016 12:35:32 INFO [218011833117] 0000299274: KSqlCursor::Execute() --> 'SUCCESS'
06/03/2016 12:35:32 INFO [218011833764] 0000299275: KSqlCursor::Close() --> 'SUCCESS'
06/03/2016 12:35:32 INFO [218011833858] 0000299276: KSql::Execute(...) --> 'SUCCESS',1 record,[sec: 115 msec: 92]
06/03/2016 12:35:32 INFO [218011833917] 0000299277: KLogin::PopulateKUAFRightsListNew(218858) --> 'SUCCESS'
[To post a comment, use the normal reply function]Topic:UAPI AddToByIDs strange behaviorForum:Content Server Development ForumContent Server:Knowledge Center

Hans Stoop

I’m wondering if support is talking about CS Projects and not about projects implemented with Binders/Cases.

Hans

From: eLink Entry: Content Server Development Forum [mailto:development@elinkkc.opentext.com]
Sent: Montag, 20. Juni 2016 17:46
To: eLink Recipient
Subject: RE Re UAPI AddToByIDs strange behavior

RE Re UAPI AddToByIDs strange behavior

Posted by Luo, Dong On 06/20/2016 11:42 AM

@Hans Stoop,

Can you elabrate more on how to achive this?

"After a binders is closed you should consider removing all roles and replace them by groups."

The answer I received from OT support is:

"After The roles themselves (Coordinator, members and Guests) can't be removed. You can only add groups to them."

[To post a comment, use the normal reply function]
Topic:	UAPI AddToByIDs strange behavior
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

David Templeton

what version of SQL server are you running? it will be at the top of the connect logs and looks like this

06/20/2016 10:55:25 INFO [252708264231] 0000000059: KSqlODBC::SubclassConnect() --> version: '12.00.4100'

David

Sachin Kalbag

Any update on how this issue was solved as we are facing same issue in prod env for CS 10.5 and the blocking transactions take 8-10 min to go away.

dluo

It has been some time and I don't remember we fixed it with a good solution. During the time, the system is upgraded to later version and the affected users were retired. As we know, new user is not affected.