Re: Message: No Server – QV11 SR2 - Page 2 - Qlik Community

rothtd · ‎2013-04-22

We have been running our environment (unchanged) for about a year and started noticing that our users were getting the red ‘Message: No Server’ on access point. We do see error messages in the Web Server logs and are working through them. The QVMC shows no issues and lists both QVS’s as fully functional. Given that this issue started appearing for seemingly no reason I wanted to ask the community if they have experienced this before.

Do you know what (at a high level) causes this error? Is it a communication error between the Web Server and QVS? [they are on the same physical server in our environment]

Is it possible that we crossed some sort of threshold like number of applications or load on QVS? (our QVS is only using a quarter of its memory, and barely any CPU).

Is it possible that a Windows patch was applied to IIS or OS and is causing this issue?

The QVWS logging level is set to ‘high’ – could the added load account for this? The log files are split daily and are < 10MB in size.

Anyone deal with this before?

Environment:

3 physical servers Win Server 2008R2, 56GB, 64-bit

QV 11 SR2 (11414)

Server 1: QVMS, QVDS, QV Directory Connector

Server 2: IIS, Directory Connector, QVS (clustered with Server 3)

Server 3: IIS, Directory Connector, QVS (clustered with Server 2)

Sample QVWS error lines:

4/22/2013 08:24:23.1782533 Error QVClient.Execute failed

4/22/2013 08:24:23.1782533 Error System.Threading.ThreadAbortException: Thread was being aborted.

at QVSWrapper.QlikViewServer.ExecuteXml(Action`2 i_Logger, String i_Command)

at QVSWrapper.QlikViewServer.GetAdminDocListForUser(Action`2 i_Logger, String i_UserName, eDocumentListType i_ListingType, eExtendedDocInfo extendedInfo, IEnumerable`1 i_Groups, Boolean GroupListIsNames)

at QlikView.AccessPoint.FileScan.<>c__DisplayClass3.<ScanQlikViewServerClusters>b__0(IQlikViewServerWrapper qvs)

at QlikView.AccessPoint.QlikviewServerSetting.Try(Func`2 f, Func`1 fail)

at QlikView.AccessPoint.FileScan.ScanQlikViewServerClusters(IEnumerable`1 servers, Action`2 logger, IUser user, Int32& liveQvsCount, Boolean respectBrowsable)

4/22/2013 08:29:52.2041806 Error System.AggregateException: One or more errors occurred. ---> System.FormatException: Input string was not in a correct format.

at System.Number.StringToNumber(String str, NumberStyles options, NumberBuffer& number, NumberFormatInfo info, Boolean parseDecimal)

at System.Number.ParseInt32(String s, NumberStyles style, NumberFormatInfo info)

at QVSWrapper.QlikViewServer.GetServerStatus(Action`2 i_Logger, String& reason)

at QlikView.AccessPoint.QlikviewServerSetting.<Ping>b__f(IQlikViewServerWrapper q)

at QlikView.AccessPoint.QlikviewServerSetting.Try(Func`2 f, Func`1 fail)

at QlikView.AccessPoint.QlikviewServerSetting.Ping()

at System.Threading.Tasks.Parallel.<>c__DisplayClassf`1.<ForWorker>b__c()

at System.Threading.Tasks.Task.InnerInvokeWithArg(Task childTask)

at System.Threading.Tasks.Task.<>c__DisplayClass7.<ExecuteSelfReplicating>b__6(Object )

--- End of inner exception stack trace ---

at System.Threading.Tasks.Task.Wait(Int32 millisecondsTimeout, CancellationToken cancellationToken)

at System.Threading.Tasks.Task.Wait()

at System.Threading.Tasks.Parallel.ForWorker[TLocal](Int32 fromInclusive, Int32 toExclusive, ParallelOptions parallelOptions, Action`1 body, Action`2 bodyWithState, Func`4 bodyWithLocal, Func`1 localInit, Action`1 localFinally)

at System.Threading.Tasks.Parallel.ForEachWorker[TSource,TLocal](IEnumerable`1 source, ParallelOptions parallelOptions, Action`1 body, Action`2 bodyWithState, Action`3 bodyWithStateAndIndex, Func`4 bodyWithStateAndLocal, Func`5 bodyWithEverything, Func`1 localInit, Action`1 localFinally)

at System.Threading.Tasks.Parallel.ForEach[TSource](IEnumerable`1 source, Action`1 body)

at QlikView.AccessPoint.QlikViewServerStatus.PingThread()

---> (Inner Exception #0) System.FormatException: Input string was not in a correct format.

at System.Number.StringToNumber(String str, NumberStyles options, NumberBuffer& number, NumberFormatInfo info, Boolean parseDecimal)

at System.Number.ParseInt32(String s, NumberStyles style, NumberFormatInfo info)

at QVSWrapper.QlikViewServer.GetServerStatus(Action`2 i_Logger, String& reason)

at QlikView.AccessPoint.QlikviewServerSetting.<Ping>b__f(IQlikViewServerWrapper q)

at QlikView.AccessPoint.QlikviewServerSetting.Try(Func`2 f, Func`1 fail)

at QlikView.AccessPoint.QlikviewServerSetting.Ping()

at System.Threading.Tasks.Parallel.<>c__DisplayClassf`1.<ForWorker>b__c()

at System.Threading.Tasks.Task.InnerInvokeWithArg(Task childTask)

at System.Threading.Tasks.Task.<>c__DisplayClass7.<ExecuteSelfReplicating>b__6(Object )<---

Thanks in advance for any input!

Report Inappropriate Content · ‎2013-05-22

Hello Rothtd,

I am very interesting by your progress on this issue... because we are currently facing exactly the same problem. 'No Server' message apprears from time to time (no pattern) with QVClient.Execute failed error message in the QWVS log

22/05/2013 13:12:45.1897056

Error

System.Threading.ThreadAbortException: Thread was being aborted.

at QVSWrapper.QlikViewServer.ExecuteXml(Action`2 i_Logger, String i_Command)

at QVSWrapper.QlikViewServer.GetAdminDocListForUser(Action`2 i_Logger, String i_UserName, eDocumentListType i_ListingType, eExtendedDocInfo extendedInfo, IEnumerable`1 i_Groups, Boolean GroupListIsNames)

at QlikView.AccessPoint.FileScan.<>c__DisplayClass3.<ScanQlikViewServerClusters>b__0(IQlikViewServerWrapper qvs)

at QlikView.AccessPoint.QlikviewServerSetting.Try(Func`2 f, Func`1 fail)

at QlikView.AccessPoint.FileScan.ScanQlikViewServerClusters(IEnumerable`1 servers, Action`2 logger, IUser user, Int32& liveQvsCount, Boolean respectBrowsable)

22/05/2013 13:13:45.6775645

Error

QVClient.Execute failed

Server 1: QDS, IIS, QMC

Server 2: QVS, DSC

Server 3: QVS, DSC

Antivirus is off

We are using Ajax only no plugin at all, so I am surprised by your comment #9

Rajesh4 · ‎2013-05-22

Anybody found a solution for this issue? We were facing this issue when we were using a cluster with 4 servers on SR1. We temporarily resolved the issue by breaking the cluster apart and have only one node up and running.

rothtd · ‎2013-05-22

We are still having this issue. It's very sporadic so it’s difficult to troubleshoot. We received a tip from support to make a memory heap setting change on our QVS servers. We are doing this tonight, if it helps I’ll post the change. We did notice that our QVS servers are not utilizing much memory/cpu (which is surprising). We are still attempting to pull all the old plug-in versions off workstations in our environment and ensure everyone is on the current version. Not much to report – no real progress on a resolution.

It’s interesting that ocourtecuisse is having the issue without the plug-in. I wonder what we all have in common? Maybe a OS patch? Or the fact that we all have a clustered QVS environment? Cynosureraj removed the cluster and resolved the issue. Our issue does seem to be worse in the morning, which is when our QVDs are refreshing – and lots of I/O is happening to the file share where our PGO files are stored (which of course are used for clustering). I know QV demands instantaneous access to this files at all times – I wonder if this might be the issue?

- Trevor

rothtd · ‎2013-05-22

QT support offered to move my QVMS onto a separate server (currently it’s on the same physical server as QVDS) as an attempt to resolve this issue. Do any of you have QVMS on it’s own server – and are you experiencing this issue?

If this were to be the culprit it would make sense why the issue would appear more frequently (and worse) in the morning, since QVDS is heavily active in the morning and only lightly active in the afternoon. I have noticed that refreshing the browser will sometimes cause AccessPoint to not display the ‘No Server’ message, or even after waiting ~10 minutes it will go away. Sometimes, this problem is ‘worse’ and we cannot get it to go away without restarting our QVS servers. Obviously, load balancing is playing a role here (my point about refreshing the browser) but my observation about it being less severe or more severe at times still stands. This is very odd!

- Trevor

Rajesh4 · ‎2013-05-23

Trevor

We spent few weeks on debugging this issue but still do not know the root cause. We have QVMS installed on the same server as QVS. QDS is installed on a separate servers. We use Ajax client. We were able to reproduce the issue without network load balancer and not seen the issue with single node env. So the issue is with Qlikview load balancing. It was a very random issue.

We were able to reproduce this issue when two are three users accessing the environment so it had nothing to do with load or hardware resources constraints. Just hitting refresh on the access point home would some times throw this errors some times it shows the documents.

It started when a .net patch was pushed unto the server hosting QVMS and QVS which was failed. Other than that we did not find anything that changed. Unfortunately repairing .net and installing the patches again did not resolve the issue. One reason could be the number of .qvw in your environment, did it go up drastically when you started seeing this issue?

Sent from my iPhone

Raj

Report Inappropriate Content · ‎2013-05-24

Hi Trevor,

Did you apply the memory heap changes and if so did they make any difference? We are also experiencing a similar issue after upgrading from 10 SR 4 to 11 SR2. (It is much more prevalent in the morning and seems to stabilise once all of the reloads are complete).

We do not have a clustered environment. All server services are running on a single machine. Our mounted folders are on a SAN. Perhaps this is something both of our environments might have in common? We do have some users still on the Qlikview 10 plugin.

Kind Regards,

Footsie

rothtd · ‎2013-05-24

footsie: We did not experience the issue in our first day since the change – but the issue was very sporadic, so this does not mean much. If the issue does not reappear in the next 2-4 weeks, then I think we can say this change helped resolve our issue. So, unfortunately, I don’t have any great news to report yet.

Keep in mind that we are actively upgrading all our plug-in versions and removing old ones (Version 9) from the environment. Another thing to add is that the server running QVDS/QVMS is quite under powered. I think this is significant because we are seeing this issue occur more frequently in the morning (when our refreshes occur). So – question for those of you following this thread: Are your QVDS server/s under a lot of load when this problem occurs (and are they properly sized to handle the load)? You can configure QVS to leave a CPU-core open for ‘other’ processes, but I don’t think you can do that with QVDS. This is why I’m wondering about the performance of QVDS and any connections to our problem (especially since our QVDS is on the same server as QVMS).

Links to the ‘desktop heap’ change that we made to all three servers in our environment:

http://michaelandqlikview.blogspot.com/2011/04/why-do-our-overnight-batches-sometimes.html

http://blogs.msdn.com/b/ntdebugging/archive/2007/07/05/desktop-heap-part-2.aspx

http://blogs.msdn.com/b/ntdebugging/archive/2007/01/04/desktop-heap-overview.aspx

footsie: We also have our mounted folders on a Windows network file share (ours is not a SAN technically). FYI – I’ve been told by QT multiple times about some issues/limitations/warnings when using SAN technology. If you are using a SAN make sure you are aware of these, and if they apply to your situation or not. This isn’t my area of expertise, so I’m not going to comment further.

Thanks!

Trevor

Report Inappropriate Content · ‎2013-05-24

@all,

We have done some progress on the issue.

After attaching windebbuger to the QVS, Qlikview support has identified that 2 QVS sub-processes for renaming/moving files were stucked. These are related to distribution.

Maybe we hold here the common point to all of us. We distribute against the QVS cluster, so the QVS is generating the temp file and then renaming it in place of the former user doc application located on a shared folder (Publisher server).

Qlikview support is suspecting the Antivirus to slow down and freeze these two QVS processes and it is true that on our servers we still have Symantec end point protection Network threat protection active (Antivirus and proactive threat protection are off)

Best regards,

Olivier

@ rothtd, when your No server message is appearing do your QVS are both up and running or one of them is crashing

rothtd · ‎2013-05-24

@ocourtecuisse - when the 'No Server' message appears all over our services appear fully operational. The QMC displays all services (QVS, etc.) as 'green' or active. Windows all shows all services as 'started'. We did notice that a restart seems to be the temporary fix, so it's possible that QVS is having problems, but QMC is still listing it as fully operational.

Miguel_Angel_Baeyens · ‎2013-05-25

Hi,

I have run into this one lately, and it turned out to be some permissions plus DNS issues between the computer running the QlikView Services and the Server storing the Shared Folder. We checked this changing in the QMC all server names for all services to the server IP address (as I mentioned in my first reply).

In this case, it was solved adding the computer IP address and name as well as those correspoding to the other services to the "hosts" file in Windows.

As a side note, in one of the servers this happened, resources were almost exhausted (RAM and hard disk). Doing some cleaning helped to make the message go away.

Hope that helps.

Miguel