[doug]

Cory determined that the issue appears to be resolved after manually the running Windows Update GUI on the problematic machines, installing the updates, and rebooting. After that, BatchPatch once again worked normally.

At this time, it’s not clear exactly what the cause of the problem was/is, but I will update this thread if/when there is more to report.

-Doug

[doug]

Cory – I’m going to email you so that we can iterate more easily. Once we’ve gotten to the bottom of the problem I will post an update here in the forum.

-Doug

[doug]

Cory – When you say “no longer wants to check for updates” what do you mean? Could you please provide details?

Is there an error? If yes, what is the exact error text?

Additionally, do you see a BatchPatchError.log file on the target machine in the remote working directory? (this directory defaults to C:Program FilesBatchPatch, if you haven’t modified it)

What happens if you run the Windows Update GUI (from control panel) manually on one of the problematic machines? Does this work as expected?

The more information you can provide, the more likely I’ll be able to help you. At the moment, however, I simply don’t have enough information to go on. There is no reason that I am aware of why things would all of a sudden stop working on your 2003 machines. Nothing changed in the BatchPatch code that would prevent it from working, so I suspect there is something that changed in your environment.

What version of BatchPatch are you using? (Help > About)

Are you using WSUS?

Thanks,

Doug

[doug]

This is odd/unexpected. My guess is that the termination itself might not have completed and that BatchPatchRemoteAgent.exe is actually still running on the target. The reason I say this is because I can’t think of another reason why there would be an access problem all of a sudden. It seems like the Remote Agent is still sitting there and Windows has it open/running, so it can’t be overwritten. This is similar to the behavior that I would expect if you tried to delete an open Word document, for example. I would start by rebooting the target machine and see if that does the trick. Let me know how it goes.

I know you said other actions appear to work, but if that doesn’t fix it, could you please test “Services > List all services” and let me know if that also doesn’t work, and if so what is the error message?

Thanks,

Doug

[doug]

Hi Stephen – This will be fixed in the next release.

Thanks,

Doug

[doug]

nolant – Yes, you can simply disable a row (or rows) that you don’t want to execute. Actions (or right-click) > Disable/enable row(s). After you’ve disabled the rows, wait X seconds, where X is the number of seconds you’ve set the Row Execution Interval to be, and all will be skipped. When the execution thread reaches a disabled row, it immediately moves to the next row (instead of waiting for the Row Execution Interval time to pass), so if row 1 is executing and you disable rows 2-10, once you wait the 30 seconds (or whatever the REI is), all disabled rows will be looked at and skipped immediately.

-Doug

[doug]

Vince – We’re working on improving error codes, and we generally try to put descriptive text next to errors whenever possible. In this case, the issue is just that you need quotes around you query, so exit code 1 is signifying Invalid Syntax:

So, for this to work for you now, try: “HKLMSOFTWAREMicrosoftInternet ExplorerSetup”

We will get this fixed in the next build so that quotes won’t be necessary. This was an oversight in the code.

Thanks for making us aware.

-Doug

[doug]

Tony – Great news is that the latest version of PsExec (v2.1 published in March 2014) actually sends all data over the network via an encrypted channel, so this is no longer an issue!

http://msdn.microsoft.com/en-us/library/bb897553.aspx

-Doug

[doug]

Thanks for the feedback and suggestions. At the moment we don’t anticipate any integration with Royal TS.

-Doug

[doug]

Excellent. Glad you got it working. Thanks for letting me know.

-Doug

[doug]

vinceco – I’m sorry that you’re having problems with importing old .bps files. There were some changes made to the 2014.5.2.9.48 version that could have caused this. In any case, if you simply create a brand new .bps file from scratch with this new version of BatchPatch, those problems should disappear.

As for ways that you can share lists of computers, there are a couple:

1. You can use .bpt files instead of .bps files. With the new BatchPatch Template (.bpt) format, you can create a single .bpt file and then load it multiple times to create multiple identical tabs in an instance of BatchPatch. BatchPatch won’t let you save over a template file, so when you click the save button you’ll be prompted to save a new .bps file for the grid. You can create your own .bpt files by simply renaming existing .bps files, or you can use the File > Generate template file (.bpt) menu option to create a .bpt file from a BatchPatch grid.

2. You can use simple .txt files. If you drag and drop a .txt file onto BatchPatch, it will load those hosts into a BatchPatch grid.

I hope this helps. Let me know if you have any other questions or problems.

Thanks,

Doug

[doug]

Perhaps you could explain to me why you believe that you need to maintain multiple copies of .bps files? I don’t think I understand exactly how you are using or how you want to use BatchPatch. Using the same Word document example, I would ask you why would more than one person edit a document at the same time? That wouldn’t make sense, right? Only one person would edit a document at one time. And so for BatchPatch, could you please explain how or why you would need to have multiple people using the same .bps file at the same time? Maybe if you could give me a better sense of how you envision using the software, I can better understand how to assist.

Thanks,

Doug

[doug]

“RPC unavailable” is almost definitely due to a firewall problem or some kind of similar access issue. Many VPN clients have firewall software built-in, which could cause this. But it could also be caused by your VPN infrastructure. In either case, the issue is that BatchPatch cannot communicate with the RPC server on the target host because it appears to not be available. Since you’ve already established that things work normally within your network, you know that the RPC server on the target is enabled. The issue is only when you’re trying to work across the VPN. That said, you need to examine your VPN setup and figure out how to allow RPC communication (and full PC to PC communication) across the VPN.

-Doug

[doug]

Hi David – I have emailed you directly to troubleshoot.

Thanks,

Doug

[doug]

Just as two or more people cannot edit a Word document at the same time from different computers, two or more people should not try to simultaneously use the same .bps files. You can use .bpt files, which are BatchPatch Template files to allow multiple users to load the same content into their own BatchPatch instances. But if/when they try to save those grids, they will be prompted to save them as new .bps files.

I hope this helps.

-Doug

[doug]

Generally speaking if the computer is removed from the network in the middle of an operation, nothing bad is going to happen. Depending on exactly what it was doing, it might complete the operation or it might not. If there is a particular update that is installing when the computer is disconnected, that update is generally going to complete installing. The rest of the updates queued to install may or may not install, depending. If a user were to manually power down his machine by holding in the power button in the middle of an installation, that would NOT be good. However, if a user simply disconnected from the network everything is generally going to be fine. You might have to finish the download/installation at a later date, or you may not have to.

-Doug

[doug]

These features have been added in the 2014.5.2 build.

-Doug

[doug]

OK – I see now that you’re saying that BatchPatch is giving you “error -2” when you try to substitute paexec for psexec. I had thought that you meant that when you ran paexec at the cmd prompt that it would give error -2.

I just installed a fresh 2012 (non-R2) domain controller (I just built it from scratch with nothing but AD and DNS), and I am able to use it as a BatchPatch target without any problems. It works flawlessly. I’ll test a 2012 R2 domain controller next to see what happens, but I suspect it will behave properly just as as all of my other 2012 R2 (non-dc) tests have. I’ll let you know. I still believe that the issue you are experiencing is unique to your particular domain controller.

I’m going to email you directly regarding the error -2 because that is a separate issue that has to do with the BatchPatch configuration for paexec.

-Doug

[doug]

It’s probably the same error just being reported slightly differently, especially if you notice that the paexec service doesn’t get removed properly either. I wonder if there is some sort of 2012 DC incompatibility with these apps. While I’ve certainly tested all Windows OSes, I have not specifically tested a 2012 DC, so I’ll see if I can reproduce on a fresh test 2012 DC.

-Doug

[doug]

I thought you were only seeing it on your domain controller and not on any other computers? I have so far not been able to reproduce it. What installation media are you using? Is this the English version of Windows? Are you only seeing it on Windows 2012 (non R2) ? What roles are you installing?

One thing you might try as a test is download the program called “paexec” which is an alternative to PsExec and is also free. Replace psexec with paexec (and rename paexec to psexec) and see if it has the same problem.

-Doug

[doug]

Hi Michael – all I meant when I said I’m not sure that it’s a “bug” per se is that it clearly only happens on a very limited number of computers. A “bug” implies that there is a problem with the code that prevents it from working properly, and that the code could somehow be “fixed” to prevent the problem. However, I tend to think that since it happens so rarely in the world of computers, that it might not be a problem with the code, but rather there might be something with the particular computer that causes the issue. This is just speculation on my part since of course I’ve never seen the code for PsExec.

-Doug

[doug]

No worries. Unfortunately there is only one output column. Your best bet is to probably create two rows per server. Sorry about that.

-Doug

[doug]

Do you have a SQL admin team who has access to all the SQL servers? A good db administrator should be able to collect all his/her SQL instance versions, using the “select @@version” in under 5 minutes. S/he wouldn’t even need to write a script because SQL management studio allows you to query multiple SQL servers at one time with just a single query. So maybe you can just ask the SQL admin team to give you this info? Would save you some time so that you don’t have to script it out. Just a thought.

Good luck!

-Doug

[doug]

Jeremy – Domain admin permissions is not sufficient to have permissions to the SQL server. I think in SQL 2000/2005 it *was* sufficient but my recollection is that starting with SQL 2008 things changed and unless the domain admins group has been explicitly added to the SQL server logins, you wouldn’t automatically have permissions. That said, it sounds like the first step will be to get yourself setup with an account that has permission to the SQL servers. Once you have that, getting the version information shouldn’t be a problem. I would then just use SQLCMD at the command line (not in BatchPatch) to test that the permissions are working properly, and then once you have it working with 1 server you can duplicate across all the servers and use BatchPatch to grab the version info. Am I missing something? Do you have a separate team who manages your SQL machines? Maybe you’ll need to work with them to give you the access you need?

I guess if you simply are not able to get access to the SQL servers, then your boss shouldn’t be tasking you with getting a list of SQL versions, no? 🙂 One other option, I suppose, would be to write a script to get the file properties on the sql exe file as described here, which wouldn’t require access to the SQL servers and would only require access to Windows: http://www.mssqltips.com/sqlservertip/1140/how-to-tell-what-sql-server-version-you-are-running/

-Doug

[doug]

Glad to be able to help, Jeremy. Let me know how it works out for you.

-Doug

[doug]

Michael – next time when the problem manifests, can you check to see if the services console (start > run > services.msc) still contains the psexesvc service? Normally what should happen is psexec will automatically clean up after itself, and the psexesvc.exe process along with the psexesvc service will both disappear automatically. However, if the issue is that it’s failing to clean up after itself, then we should see psexesvc service still exist in the services.msc console. If that’s the case, instead of rebooting you can try running on the domain controller the following commands to manually clean up the service. See if that works and if you can successfully use BatchPatch after you do that.

sc.exe stop psexesvc

sc.exe delete psexesvc

(or I think you should be able to run this remotely with “sc \targetDC delete psexesvc”

You can see a bit more info at this forum: http://forum.sysinternals.com/delete-psexecservice-on-the-remote-computer_topic10926.html

However, I don’t know if it’s safe to call this a “bug” with PsExec. It seems to be more of a failure, that may occur in certain cases. I don’t know what those cases are. It clearly only happens in certain situations, such as in your case on just your DC sometimes. Personally, I have never experienced it in testing or in production, and I use PsExec a LOT, on thousands of computers. Overall, considering how widely PsExec is used, I think it’s probably a relatively rare situation. That said, considering that it only happens on one of your machines, and that in the world of IT it’s probably relatively rare seeing as though PsExec gets TONS of successful use all over the world, I would still think that there is likely *something* happening on that DC that is somehow not compatible. I don’t know what it could be or why it would continue to happen intermittently on that one computer in your environment. Unfortunately I don’t think we are going to see a fix or code change in PsExec that will address the issue.

Anyway, I would definitely be curious to know if using the SC.exe commands works for you. I realize it’s not a solution for the cause of the problem and is instead just a fix/workaround, but it would still be helpful to know if that works in your situation.

I’m sorry that I don’t have any better thoughts or ideas on the subject.

-Doug

[doug]

OK – It’s hard to say exactly what is happening here. What we SHOULD see is when the BatchPatchRemoteAgent.exe starts executing, it creates a number of .log files. However, those files are not getting created, which indicates that it is not getting executed properly in the first place. The only other place where there could be a problem is with PsExec. When execution begins you should see in Task Manager list of processes the PSEXESVC.exe. Perhaps that is what is actually getting killed by your AV software or by something else on the machine?

You could try a test to run at the cmd prompt on the BatchPatch computer: “psexec \targetDomainController notepad.exe”

This should spawn a notepad.exe process (hidden) on the target computer. In Task Manager on the target computer we should see both PSEXESVC.exe and notepad.exe. They would both stay there indefinitely or until you manually kill notepad.exe. After you kill notepad.exe, the PSEXESVC.exe process should disappear on its own.

In any case, given the fact that all of your other computers work, and given the fact that this DC does sometimes work right after it is rebooted leads me to believe that the issue is still something specific to the DC. I don’t think this is an issue with BatchPatch per se. Perhaps you can try to evaluate what else on the DC is different from your other computers? Is there anything different at all? Is there anything that happens on the dc after some period of time? Since BatchPatch works properly after a fresh reboot, does something change on the DC after it is online for a certain period of time?

There is no debug mode for the remote agent, but it doesn’t seem like it’s crashing. As you pointed out, if it were crashing, we would expect to see something in the event log to indicate the crash. It seems like it’s somehow getting killed by the DC. To date, we have zero reports of the remote agent ever crashing for any customers. The code for that exe is very clean, and this really seems like something that the DC is doing. I just don’t know what or why. I’m sorry I don’t have any better suggestions at the moment.

Let me know how it goes.

-Doug

[doug]

Actually – here you go. Here’s a VBScript that you can use to retrieve the CPU info in one shot without needing to manually multiply anything:

http://batchpatch.com/downloads/CPUCoreCount.vbs

In BatchPatch use “Local Command” with the following syntax:

cscript C:scriptsFolderCPUCoreCount.vbs $computer

[doug]

Hey Jeremy – We actually have on the list to add some basic SQL query capabilities, but it hasn’t been added yet. In the meantime, this will work for you. Create a “Local Command” in BatchPatch as follows:

sqlcmd -S $computer -d master -Q "select @@version" -U username -P password

---

With regard to CPU socket/core count, this might work for you. Off the top of my head this would really need to be done in a loop in order to do the multiplication for you, so you could write a VBScript or Powershell script to do that for you, but since I don’t have time right now to pull together a proper script, you can use the following 2 commands to give you the info you need. But you’d have to manually multiply the NumberOfProcessors by the NumberOfCores to get the total physical core count in the machine. You can eliminate the NumberOfLogicalProcessors from the command below, but if you leave it in it will essentially tell you if HyperThreading is enabled or disabled for a machine. If HyperThreading is enabled, then you’ll see that the NumberOfLogicalProcessors will be double the NumberOfCores. If HyperThreading is NOT enabled, then the NumberOfLogicalProcessors will be equal to the NumberOfCores. It might make most sense to create 2 rows per server, so that you can get the results for both of these queries, and then you can export to HTML, copy and paste the grid to Excel, or save to a .bps file and then import the .bps file as .xml into Excel:

CPU Socket Count:

WMIC PATH Win32_ComputerSystem GET NumberOfProcessors

CPU Core and Logical Processor Count:

WMIC PATH Win32_Processor GET Name,NumberOfCores,NumberOfLogicalProcessors

---

With regard to he window not responding, I’m not sure exactly what to make of it. Depends on how it’s not responding. If there is a lot of data in each cell, that will slow down the grid’s click/scroll performance (we have a fix for this coming in the next build). If that’s what you’re seeing, then your best bet is to close the columns that contain the large amounts of data or clear their contents. This could be the “All Messages” column if you’ve run many commands, or if you have the Remote Agent Log column open, that one typically gets pretty big, so closing that would help.

If the lack of responsiveness is not related to clicking and scrolling around in the grid and is actually just waiting for queries to return results, then you probably should try running less at once. I’ve never tried to query 1671 machines at once, so I’m not sure what to expect. However, I would expect that a couple hundred at a time should perform fine. If you really get stuck you could also launch a number of instances of the BP.exe and put a couple hundred in each instance. Again, it really depends on the nature of the slow-down.

I hope this helps.

-Doug

[doug]

Do you have any kind of anti-virus or host intrusion prevention software running on this DC? It sounds like the BatchPatchRemoteAgent.exe is getting deleted on the DC before BatchPatch is able to execute it, and anti-virus or host intrusion prevention software would be the likely culprits in this case.

The normal process that BatchPatch follows is:

1. Connect to remote computer

2. Put BatchPatchRemoteAgent.exe on remote computer

3. Remotely execute BatchPatchRemoteAgent.exe and retrieve its output

4. Delete BatchPatchRemoteAgent.exe

5. Close connection to remote computer

Exit Code 2 likely means that BatchPatchRemoteAgent.exe is being deleted by the remote computer as soon as BatchPatch puts it there, so then when BatchPatch tries to execute BatchPatchRemoteAgent.exe, the file doesn’t exist and so it throws error 2.

If you watch the C:Program FilesBatchPatch directory on the target computer when you execute the BatchPatch “Check for available updates,” you should be able to confirm that this is what’s happening. When you execute “Check for available updates” I would expect that we would see the BatchPatchRemoteAgent.exe appear for just a moment and then disappear, and then BatchPatch throws the Error 2. Is this what is happening? Are you able to determine what program on the DC is deleting the file?

I hope this helps. Please let me know what you find.

Thanks,

Doug

[Author]

Posts