SCCM Task Sequence Failed - Program failed (download failed - content mismatch)

[lee_SCCM]

Hi,

 

We are currently using Task Sequences in order to deploy software packages sometimes multiple sometimes simple packages.

 

Our site setup means we have a CENtral server 3 primary and the rest made up of Secondary or BDP's.

 

As the title suggests we seems to have this ever occuring issue where by ( for example my deployment today that prompted this)

 

I am deploying Adobe and one other application to 400 machines last night. This morning I come into around 18% of these have failed and generally the error is similar to below:

 

05/10/2010 16:11:44 Accepted Program received 10002 12177032

05/10/2010 22:59:01 Running Program started 10005 12202192

05/10/2010 23:04:06 Succeeded Program completed with success 10008 12202596

05/10/2010 23:04:07 Waiting Waiting for content 10035 12202606

05/10/2010 23:04:29 Failed Program failed (download failed - content mismatch) 10057 12202849

06/10/2010 05:00:01 Expired Program expired 10019 12207830

 

 

As you can see highlights the download failed content-mismatch is my issue,.

 

Now what is weird is that a large proportion work perfectly fine however 18% is to big a number to fail on a regular basis.

 

I have searched the length and breath of the internet for answers and thet vary in how to potentially resolve it.

 

I have tried updating the DP but that did not work, i have not turned of binary replication as its a BDP, im reluctant to do anything ultra extreme baring in mind that so mnay complete with no issues.

 

In this TS there are two apps and on a few machines i have sampled the CAS.log file it seems to reference one or the other package that failed to downloard and sure as anything on that machine it is not in the cache folder.

 

 

 

Anyone throw anylight on this be most apprecaited.

 

Thanks in advance

[lee_SCCM]

Hi,

 

We are currently using Task Sequences in order to deploy software packages sometimes multiple sometimes simple packages.

 

Our site setup means we have a CENtral server 3 primary and the rest made up of Secondary or BDP's.

 

As the title suggests we seems to have this ever occuring issue where by ( for example my deployment today that prompted this)

 

I am deploying Adobe and one other application to 400 machines last night. This morning I come into around 18% of these have failed and generally the error is similar to below:

 

05/10/2010 16:11:44 Accepted Program received 10002 12177032

05/10/2010 22:59:01 Running Program started 10005 12202192

05/10/2010 23:04:06 Succeeded Program completed with success 10008 12202596

05/10/2010 23:04:07 Waiting Waiting for content 10035 12202606

05/10/2010 23:04:29 Failed Program failed (download failed - content mismatch) 10057 12202849

06/10/2010 05:00:01 Expired Program expired 10019 12207830

 

 

As you can see highlights the download failed content-mismatch is my issue,.

 

Now what is weird is that a large proportion work perfectly fine however 18% is to big a number to fail on a regular basis.

 

I have searched the length and breath of the internet for answers and thet vary in how to potentially resolve it.

 

I have tried updating the DP but that did not work, i have not turned of binary replication as its a BDP, im reluctant to do anything ultra extreme baring in mind that so mnay complete with no issues.

 

In this TS there are two apps and on a few machines i have sampled the CAS.log file it seems to reference one or the other package that failed to downloard and sure as anything on that machine it is not in the cache folder.

 

 

 

Anyone throw anylight on this be most apprecaited.

 

Thanks in advance

[Peter van der Woude]

Hi,

 

We are currently using Task Sequences in order to deploy software packages sometimes multiple sometimes simple packages.

 

Our site setup means we have a CENtral server 3 primary and the rest made up of Secondary or BDP's.

 

As the title suggests we seems to have this ever occuring issue where by ( for example my deployment today that prompted this)

 

I am deploying Adobe and one other application to 400 machines last night. This morning I come into around 18% of these have failed and generally the error is similar to below:

 

05/10/2010 16:11:44 Accepted Program received 10002 12177032

05/10/2010 22:59:01 Running Program started 10005 12202192

05/10/2010 23:04:06 Succeeded Program completed with success 10008 12202596

05/10/2010 23:04:07 Waiting Waiting for content 10035 12202606

05/10/2010 23:04:29 Failed Program failed (download failed - content mismatch) 10057 12202849

06/10/2010 05:00:01 Expired Program expired 10019 12207830

 

 

As you can see highlights the download failed content-mismatch is my issue,.

 

Now what is weird is that a large proportion work perfectly fine however 18% is to big a number to fail on a regular basis.

 

I have searched the length and breath of the internet for answers and thet vary in how to potentially resolve it.

 

I have tried updating the DP but that did not work, i have not turned of binary replication as its a BDP, im reluctant to do anything ultra extreme baring in mind that so mnay complete with no issues.

 

In this TS there are two apps and on a few machines i have sampled the CAS.log file it seems to reference one or the other package that failed to downloard and sure as anything on that machine it is not in the cache folder.

 

 

 

Anyone throw anylight on this be most apprecaited.

 

Thanks in advance

 

Lets start with, that I think that having multiple BDPs is a good reason to enable Binary Differential Replication (BDR). As BDR is used to update package source files with a minimum of additional network traffic (take a look here: http://www.petervanderwoude.nl/post/ConfigMgr-2007-and-Binary-Differential-Replication.aspx).

 

Now back to the issue, the machines that fail what do they have in common? Same DP? Same ...??

[Peter van der Woude]

Also, I merged both topics into one, as it was double posted.

[lee_SCCM]

Lets start with, that I think that having multiple BDPs is a good reason to enable Binary Differential Replication (BDR). As BDR is used to update package source files with a minimum of additional network traffic (take a look here: http://www.petervanderwoude.nl/post/ConfigMgr-2007-and-Binary-Differential-Replication.aspx).

 

Now back to the issue, the machines that fail what do they have in common? Same DP? Same ...??

 

 

Peter,

 

Thanks for your response,

 

I am doing a european wide deployment multiple offices that all have seperate BDP's im please to have some re-assurance that it is not the binary potentially causing the issue.

 

In terms of what they have in common from my group of machines that I have there are a few offices which have produced the same problem they would have perhaps 2 BDP's at that site or they would go to the european primary for the package

 

I have enclosed 2 different screen shots of the Log ( with trace so its easier to see ) of the process it seems to go through as it bombs out.

 

If there is anything specific you want me to check in common let me know

 

Thanks

 

Lee

 

Log without BDP.bmpLogwitha_BDP.doc

[Peter van der Woude]

A hash mismatch is almost always a difference in the Package Source version. Did you check if all the Distribution Points have the same Package Source version of the Packages? If so, just do a Package Refresh on the problematic DPs (see also: http://technet.microsoft.com/en-us/library/bb892806.aspx).

[lee_SCCM]

A hash mismatch is almost always a difference in the Package Source version. Did you check if all the Distribution Points have the same Package Source version of the Packages? If so, just do a Package Refresh on the problematic DPs (see also: http://technet.microsoft.com/en-us/library/bb892806.aspx).

 

 

Ywah I have updated the DP several times but to no avail, making it a version 2 didnt fix it either. Why would this affect some not others, like i say when WOL is working i have around a 72% success rate ( aside from minor anomlies) just that is not good enough.

 

only thing i have not done is tore it down from the DP let it refresh then push it back up again fresh, think it would make any difference.

 

Also its soooooo sporadic, my london site it never happens on :S rattling my brains on this one!

[Peter van der Woude]

In terms of what they have in common from my group of machines that I have there are a few offices which have produced the same problem they would have perhaps 2 BDP's at that site or they would go to the european primary for the package

But does this mean that there a few BDPs that fail to all the clients that are trying to download?

[lee_SCCM]

But does this mean that there a few BDPs that fail to all the clients that are trying to download?

 

 

no as for example one BDP has a large chunk of clients that recieved the package fine and it was a success on the report. That site there are two BPD's those clients worked whereas some failed like the one in that particular log file.

 

Why would some work and some not? makes no sense if the issue is the package and the hash vaule :S

[Peter van der Woude]

Do these clients even download the package?

[lee_SCCM]

Do these clients even download the package?

 

 

No, as I have in the log file, for example this TS there are two packages, in most cases it is either one or the other that fails. in some instances one of them is in the cache folder so it does not require to re-download but where it has to, it seems that it knows where to get it from but then the hash value is wrong so it bombs out.

 

This in turn bombs out the sequence and the software is never applied.

 

Im going to throw a spanner in the works so all my cards are on the table here. Now a product called NOMAD is used at what seems to be the BDP sites, not sure if your aware of it but its a peer-to-peer tool that 'from what it says on the tin' can help out get round the BDP's 10 connection maximum isue however my next stage of testing is to determine if in some way this could interfere at all????

 

Im told its certified to work with SCCM but just so you know the environment and you font go down an alley when i have not explained all the setup

 

Thanks