Posts made by atripp

I totally agree! Thanks for pointing this out.

I've filed this is a bug report (OT-254), so it will get fixed in a future maintenance release

You can use $PSEval for this pretty easily:

set $a = 2;
set $b = 3;
set $c = $PSEval($a + $b);

You can do this with default values.

template Check_Version<$App, $Role=default>
{

}

And then test if it's the default value to see if it wasn't passed in

Soon --- this is a known issue, but we are in the process of resolving it; please see InedoCore#96

This is a known issue, but we are in the process of resolving it; please see InedoCore#96

At this time, the only thing we're aware of that can cause "Could not resolve function" is that the function has not been loaded by the service. Functions are stored in extensions, and Split is stored in Inedo Core extension, so this would actually cause all sorts of errors.

I created an issue (BM-3255) to give a warning if this extension is not loaded. But please restart the service, check for extension load errors in the message center, and try again.

Restarting the service would be the only way to resolve this, since extensions are loaded at service start.

The error message is coming from within the network stack inside of ASP.NET's hosting components, and errors like this tend to indicate problematic underlying infrastructure, such as drivers, hardware/memory. They are often quite rare and not reproducible, and thus can be difficult to diagnose or debug.

The Integrated Web Server was not designed to handle problematic infrastructure, and as such will simply crash when receiving certain "impossible" scenarios from the network stack. Certain errors are not safe to try/catch (because they imply a corrupt process memory state), which is why they are done.

If you are finding "service autorecovery" not enough, then you should switch to IIS hosting. IIS is very robust, and uses a worker process model and can handle a lot of these underlying problems. When a worker process crashes (which it will do in the same scenario -- unsafe exceptions), IIS will spin up a new one almost instantly.

This was a result of a imcompatable script with 2005, but it has been fixed as of 5.1.11. We are also displaying warning if you're using unsupported version of SQL

We are only aware of that message occurring in the two cases I mentioned, and haven't had reports of the contrary. From here, please use a tool like Fiddler or Wireshark to identify the exact request/response patterns that are happening. You can see the authentication headers sent, and it's possible there's a typo or something in a name or password.

We take problems very seriously, but we need to prioritize where to spend engineering resources tracking down problems. This was a edge case that we could not reproduce, and had a very minor impact, and only affected a very small number of free users with a very specific configuration; a free user was able to track down the underlying cause and we fixed it.

Hi,

the script who is stored in a global module

##AH:UseTextMode
module ParseURL<out $JsonFile>

{

set $URL = $DEPLOYABLE_URL;

set $PATH = $DEPLOYABLE_PATH;

set @MYLIST = @();

Log-Debug Split URL /;

foreach $DATA in @Split($URL,"/")
{
    # Log-Debug $Data;
    set $DATA = $Trim($DATA);

    if $DATA != ""
    {
        set @MYLIST = @ListInsert(@MYLIST, $DATA);
    }
}

......

Sometimes, it works, but the major part of the time I have the error resumed above.
I found nothing in the diagnostic center et nothing in the event viewer of the widows server.

Does exist other logs somewhere ?

Kind Regards

You can reset the Admin user account password by stopping the ProGet Windows service, then running ProGet.Service.exe and selecting the "ResetAdminPassword" option. Make sure to restart the Windows service when finished.

Can you give us a little more context? Like, what is your OtterScript overall?

One possibility is that the INedoCore extension had a load error, though I suspect you'd have another problem if that were the case.

Hello,

I've added an ExitCode output parameter to the SHCall operation.

You can download a preview build of the extension from here: https://ci.appveyor.com/project/Inedo/inedox-linux/build/1.0.3-CI.7/artifacts

Currently assets do not have any user-definable metadata like this, I guess because that would make them a lot like packages? Assets also aren't intended to be like a "SharePoint replacement", which is why we don't want to put a lot of extra features like versioning, etc. These are also hard to replicate, etc.

That being said, we are open to changes, so please consider using the Feature Request Process. Note that, we also want to evaluate "why" and "how this would be used", and that's part of the feature/change process.

Thanks.

FYI: Retention policies should be ignored on free versions. But UNKNOWN does indicate the service user; so it's likely a retention policy job (which is run as part of feed cleanup).

You can see the logs of those, and maybe see some more info...

There is a task attribute called "Delete Package". Just make sure you have a Task that has that attribute. You can see more info about it here:

https://inedo.com/support/documentation/proget/administration/security/creating-tasks

ProGet most definitely does not "lose package" on an upgrade, so there must have been some sort of configuration change, either in the ProGet software (retention policy, package store), or external to the software (files deleted from disk).

It's highly unlikely that this configuration changed as part of the upgrade, particularly from those versions. Most likely, your database backups just happened to be dated from before the upgrade because the installer will auto-backup for you; so, the upgrade is really a red herring. You can search ProGet's audit logs to see if anyone in the software deleted them.

But regardless, if the package files are gone, and you don't have them backed up (the installer will not do this for you), then they are gone. There's nothing you can do, so I think your idea of searching for fragments from user caches, and then uploading those, is the best you can do.

NOTE: Unless you explicitly configured a retention policy for ProGet to delete files, the software will not delete these files on disk. Even when you delete a feed or uninstall the software, the package files remain. We do this for safety purposes.

Please make sure to read our backup and restore instructions.

Definitely not, they are on the low-end of the medium size... and just to be clear you shouldn't be calling dbo.Dashboards_GetDashboardInfo with more than 10 packages. It's only designed/tested for 0, 1, and 10.

Based on the table sizes and distribution, there's no reason it should be performing so slowly. Actually it works fast even on instances with 100x the packages that you have.

Maybe there's something else with your data that we can't see based on the tables you sent us. You should be able to view/edit the code of the procedure and underlying database views; perhaps you can find something by investigating the queries and data relationships?

Without seeing your log, I really don't know. But in this case...

WARN : 2018-02-21 22:37:21Z - ID: 107;	Name: BUILD-SOURCE;	Scope: Release

We have a legacy release template variable called "BULID-SOURCE". This will only show up in an application's settings page, it will not show up under global.

So how can we find the application? well, if I run this query, SELECT Scoped_Application_Id FROM VariableDeclarations WHERE VariableDeclaration_Id=107 I can find the application id. From there, I can enter it in the navigation bar, and find the variables.

Yes. But hmm ... maybe it's a disabled application? You should be able to find those in the VariableDeclarations table; those IDs refer to those.

These Legacy Template Variables (Release, Build, Execution, Promotion) are defined at the application-level, not at the system-level.

If you View Full Log, you should be able to find a block that looks something like this:

WARN : 2018-02-21 22:37:21Z - Legacy template variable declarations found:
WARN : 2018-02-21 22:37:21Z - ID: 107;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 92;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 108;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 109;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 111;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 112;	Name: BUILD-SOURCE;	Scope: Release
WARN : 2018-02-21 22:37:21Z - ID: 113;	Name: BUILD-SOURCE;	Scope: Release

If that's not enough information to locate, you can look directly in the SQL Server Database to find out.

I'm afraid I don't really understand the requirements for what you're trying to accomplish...

Here are the differences between the two extensibility points:

• A Package Filter allows all locally-stored ProGet packages to be filtered from being shown in feeds by additional logic.
• A Package Access Rule Allows for rules to block downloads of packages based on additional logic.

The Package Access Rule is used by the Whitesource extension, so you can reference how that works. We don't have any publicly available uses of Package Filters; it was incorporated for a very specific usecase with publicly-available feeds.

I'm sorry, I accidently replied to here instead of this post. Please disregard it; the issue has nothign to do with your license.

I don't have an answer to your question, but another engineer will help soon. Note that if you have a paid license you can use a support ticket.

But someone will respond to this soon!

Yes, it just involves changing the license key.

Please refer to License Key Management

I'm afraid we don't have a guide on Azure-hosted AD at the moment, but from talking with other customers, they said they used the Advanced Configuration documented here https://inedo.com/support/documentation/various/ldap/ldap-active-directory

Single sign-on is unscheduled so I don't have a timeline, but it seems quite soon from what 've heard... 2019/Q1 or even Q4.

ProGet already supports Azure-hosted AD Authentication; if you're looking for Federated Authentication (single-sign on), that support is coming across all of our products.

Yes, it just involves changing the license key.

Please refer to License Key Management

ProGet calls the procedure on those pages 10 and 1 for Packages_Count. If it's really slow at 10 packages, you may want to try rebuilding/defragmenting your SQL Server indexes.

Good questions;

1. You can configure ProGet to block packages with "unassessed" vulnerabilities; this is not the default configuration because a lot of times vulnerabilities are discovered after you've already been using a package anyways

2. That's true, but maybe we could expand webhooks? I'm not sure, but it seems like it could be a feature request if it's something you'd actually want to use, etc.

Hello; just to let you know, the Inedo ProGet Plugin for Jenkins has been updated, and may work better.
Have you had an progress onthis otherwise?

TFS labels can be a bit complicated, and the way that Visual Studio (GUI) vs tools (tf.exe, BuildMaster, etc) interact with them are different. What may be happening in the GUI is that multiple labels are being created with the same name, and the the GUI is "searching" for files by label name (as opposed to "getting" those files).

If you were to do a tf.exe get (https://docs.microsoft.com/en-us/vsts/tfvc/get-command?view=vsts), for example, you should see the same behavior. Instead, makes sure to first create a label at the $/ root level (e.g. tf label myLabel $/ ), and then you can sub-label items.

Hello; please check out this KB article for help on how to update the urls.

https://inedo.com/support/kb/1014/changing-an-inedo-products-url-with-the-integrated-web-server

Alternatively, you can use IIS, which will be easier for your web admins to maintain.

It may have been a temporary outage, or a proxy problem, or something. You can ignore it. If you restart the service it will probably go away.

Hi Sebastian,

You can access all of the DNC packages using the ODATA (v2) protocol API.

If you're finding an issue with it, it's totally unrelated to the API protocol. Please submit a different question with specific packages, reproduction steps and we can investigate

Cheers,
Alana

ProGet runs fine under a domain account.

The Inedo Hub is the desktop application used as an installation center, and doesn't impact the way ProGet is run. Next time you see an error with the Hub, please send us the error report so we can investigate it better.

This particular error (the 500) will be fixed with PG-1313

This will be fixed with PG-1313

PublishLocation is where you publish packages to, and SourceLocation is where you download packages from. It's typically the same, but sometimes it isn't.

Please don't forget to mention ProGet in your article as a good private repository ;)

A package is uniquely identified by its version, so different package versions will need different numbers. But I think you can use the pre-release labels, at least in some sort of way to do product support?

Take a look at this video from our recent Hedgehog webinar: Using Semantic Versioning to Overcome Package Deployment Challenges. You can see the whole webinar on the Hedgehog page as well.

Universal packages support SemVer2, which is a three-part version number scheme with additional labels for pre-release and build metadata.

The NuGet "4-part" scheme was proprietary, undocumented, and had very strange quirks and behaviors, which is why NuGet has deprecated it years ago.

If you absolutely need to use a proprietary/non-standardized version number, then I recommend you to just add it as metadata to the package (_myVersion or _componentVersion) .

Otter, BuildMaster, Hedgehog, and ProGet for Windows all require any supported version of Microsoft SQL Server. So, 2017 would be supported, since Microsoft supports it still. 2005, however, not.

Hello Telly,

Is your Docker feed named docker in ProGet? I get this error when I try to push to a feed that doesn't exist in my local ProGet instance.

Hello Dwight,

This should be logging an error message, but it's not. I've filed PG-1311 to fix the lack of a useful error. I've attached a modified build of ProGetCoreEx.dll to PG-1311.

If you replace ProGetCoreEx.dll and try this again, is there a warning message starting with Unable to open file in /administration/logs?

This is basically an error reporting an error; it can sometimes happen, and thus it masks the underlying error. The usual cause is the client disconnected unexpected, and the error can't be reported.

You can switch to INTEGRATED PIPELINE mode to see the full error

I'm afraid there's not any further configuration. This will happen on any server, physical or virtual, when the SQL Server cannot be contacted.

It's usually related to network connections, or the SQL Server is just totally bogged down from other things. See if rebooting fixed it, that will often help.

Since this is anew installation, why don't you totally start over (Delete the database).

Then, make sure that when you create the database, your user will be in the "dbo" schema. Then it should all work.

I see, so you're saying that that the ReSharper client appears to not work with a ODATA/v2 API? In this case, this is most certainly a bug or a misconfiguration and you should share/communicate it to JetBrains.

There is virtually no support in commercial or open source private repositories for JSONLD/v3. While it's on our roadmap to implement, it wouldn't come until late this year... and obviously it would require upgrading your PRoGet server software, which many organizations don't like doing for at least 6/12 months after a release (especially for something like ProGet).

So, this would effectively makes ReSharper unusable for private extensions.

I'm not familiar with Zabbix, but there are two things you can do to check about the status of the components (Database, Web, Service):

• If the database (SQL Server) wasn't working, or there was a problem with the website configuration, then visiting / will yield a 500 or some other not available message.

• You can just query the ProGet service run state directly.

• You could parse the text on the /connectors page or use the native API to get a status of the connectors.