Latest Contributions by DannyKrouk

@cboyerTucson  Ha!  Thanks.  Virtual cookies are almost as good as real ones! 😉 ... View more

@cboyerTucson  Thanks for the kind words.   You might be looking for: Which gives you this kind of thing: If that's not quite what you're after, perhaps you can tell me what is missing relative to your interests? Thanks 🙂 ... View more

@SGTomlins  Could we please continue this conversation by email at [email protected]?  I think you and I would benefit from an opportunity to coordinate our thinking and communication in details that I do not think are relevant to this forum.  Perhaps we need to graduate to a screen share to truly understand what is going on, I don't know.  But, this forum-post method, from my perspective, has no more gas in the tank for our purposes. ... View more

@SGTomlins  Thanks.  The query is seeing your backups.  That's good.  From appearances, the default definition of "recent" for the egdbhealth query is out of alignment with your default definition of recent.  It sees your back-ups.  They are all from a week or more ago.  If you run this tool right after you run take your backup, the results should be different.   Do you see another interpretation of this data? ... View more

@SGTomlins If you would, please run the SQL statement against your system in SSMS and show me a screen capture of the result in an email ([email protected]).  Thank you. 🙂 ... View more

@SGTomlins  Thanks for the kind words and the feedback.   This is the query that supports the backup findings: declare @min_backups int = 1 declare @max_minutes_since_backup int = 10080 -- one week declare @backups int select @backups = count(*) from msdb.dbo.backupset s where s.database_name = DB_NAME()    if (@backups < @min_backups ) select cast(@backups as nvarchar(8)) as backups, 'Critical: Fewer than ' + cast(@min_backups as nvarchar(8)) +  ' backups exist' as comment else SELECT top 10 DATEDIFF(mi, s.backup_start_date, getdate()) minutes_since_backup, s.database_name, m.physical_device_name, CAST(CAST(s.backup_size / 1000000 AS INT) AS VARCHAR(14)) + ' ' + 'MB' AS bkSize, CAST(DATEDIFF(second, s.backup_start_date, s.backup_finish_date) AS VARCHAR(4)) + ' ' + 'Seconds' TimeTaken, s.backup_start_date, CASE s.[type] WHEN 'D' THEN 'Full' WHEN 'I' THEN 'Differential' WHEN 'L' THEN 'Transaction Log' END AS BackupType, s.server_name, s.recovery_model, 'Warning: No recent backups.  Review back-up procedures relative to requirements. Age (in days): ' + cast( (DATEDIFF(mi, s.backup_start_date, getdate()) / 1440)   as nvarchar(8))   as comment FROM msdb.dbo.backupset s INNER JOIN msdb.dbo.backupmediafamily m ON s.media_set_id = m.media_set_id WHERE  s.database_name = DB_NAME()  and DATEDIFF(mi, s.backup_start_date, getdate()) > @max_minutes_since_backup ORDER BY DATEDIFF(mi, s.backup_start_date, getdate()) ASC It strikes me that there are a few possibilities that explain your outcome: 1. There is filter logic in my query which has a different idea than you about what "recent" is. 2. There is a flaw in the query logic relative to how (some more recent version of?) SQL Server records its back-up activity. 3. There is some kind of SQL Server back-up happening that would not normally be recorded in the system tables queried ... for some reason (really a variant of #2) As you may be able to tell from the query text, what we are examining are tables like msdb.dbo.backupset and msdb.dbo.backupmediafamily.  Do the back-ups that you see in your error log also appear in these tables? ... View more

@SGTomlins  Thanks for exercising the tool and sharing feedback.  I will pass it along to the product owner and developer. 🙂 ... View more

@SGTomlins  Thanks for the kind words!  So that you are aware, Esri has produced a Python-based script which does a portion of what GIS Enterprise Reporter does (but not require a non-product executable application): https://support.esri.com/en-us/knowledge-base/configuration-reporter-000033086#:~:text=The%20Configuration%20Reporter%20is%20a,1%20are%20supported.  GIS Enterprise Reporter isn't going anywhere (and I will continue to work on).  But, the Python-script will allow some of the capabilities to be used by more customers. 🙂 ... View more

What you are describing, where different clients use different CNAMEs to access the same ArcGIS Enterprise components is, as far as I know, not supported (Esri Tech Support is the authority on that). To your specific question, F5 sales reps and informed observers agree (in a rare reversal of the typical relationship between sales reps and informed observers -- ha, ha!) that F5's BIG-IP can do almost anything.  BIG-IP has been changing HTTP headers for years.  So, with considerable effort and great responsibility, you can 'middleware' your way to make ArcGIS Enterprise think that all requests come from the same CNAMEs.  But, I don't think Esri Tech Support would help you do that.  And, I would not be able to "talk you through" the principles or procedures here. ... View more

When you federate, you provide Portal with two parameters: (a) the ServicesURL and the ServerAdminURL.  The ServicesURL *must* be through the proxy (F5, which then goes to the Web Adaptor, which then goes to ArcGIS Server).  The ServerAdminUrl is up to you.  It is *private* to the system's administrative communications.  You can send it through the proxies or not. FieldMaps is one application among many ... and can be hard to study when compared to a web browser.  I would recommend that you study the problem from a web browser client.  If you cannot repro a problem with your browser, then the issue might be with FieldMaps ... or networking that is not part of your proxy configuration. ... View more

@vbort_ef  Thank you very much for the kind words.  I have no plans to stop maintaining the tool ... we use it in Esri Professional Services.  If you have ideas for improvement, or come across defects, please let me know. 🙂 ... View more

@TimWillson "It depends"? 😉 If the Portal and Server Web Adaptors are not on separate servers, then nothing about the load balancing approach would be different ... because BIG-IP is balancing the web servers, not the Esri software servers themselves. If the Web Adaptors are on separate servers, there are many design alternatives.  Maybe you are presenting them on a single VIP and using path-based routing to land on separate backend pools.  Maybe you are exposing two VIPs, each routed to its separate backend pool. I think, for there to be a specific answer, there needs to be a more detailed specific design.  And, that level of detail (in question and in answer) is a bit outside the intent of this article.  That is, this article is focused on foundational education such that you can work out the details with your team of F5 and GIS admin experts. ... View more

@AndreaB_  I hope the additional explanation is helpful to you and others.  I suspect you are not the only person with a question like this.  Regarding other resources, I confess that I've not reviewed the content myself, but this is a big new offering from Esri that seems like it would have information in this space: https://architecture.arcgis.com/en/introducing-the-arcgis-architecture-center.html?rsource=https%3A%2F%2Fwww.esri.com%2Fcontent%2Fdam%2Fesrisites%2Fen-us%2Fmedia%2Ftechnical-papers%2Farchitecting-the-arcgis-system.pdf (The Well Architected Framework). ... View more

Hello @AndreaB_  - Thanks for the question.  I am not sure how this statement in the doc, "Portal for ArcGIS does not support SSL offloading through a reverse proxy server/load balancer. Therefore, if your configuration uses a reverse proxy server, it must forward traffic to either the ArcGIS Web Adaptor or directly to Portal for ArcGIS over HTTPS" is in conflict with this blog post. The blog post describes how to forward traffic "to the ArcGIS Web Adaptor" (really to the web server hosting it).  So, that seems in alignment. The bit about "does not support SSL offloading" is, perhaps, confusing because "SSL offloading" can mean different things.  But, both F5 and Wikipedia agree with the sense that I think Esri means in its documentation.   "Offloading" does not mean "terminate and re-initiate TLS sessions".  Rather, "offloading" means "terminate TLS sessions and leave the back-end traffic unencrypted".  This post does not describe an "SSL offloading" configuration; it describes a "terminate and re-initiate" configuration ... which is what Esri supports. Am I missing something? ... View more

@MatejVrtich Thank you for the appreciation. I don't have plans to write such an article.  Much of the work would be the same.  The main difference is that you have to re-write the 3xx series response location headers whenever they contain the back-end machine name/native port combination.  This is only needed for the front-side channel, not the admin pathways.  This document notes the need to do so (but leaves the implementation up to the reverse proxy admin ... because reverse proxies all have different ways of doing such things): https://enterprise.arcgis.com/en/portal/latest/administer/windows/using-a-reverse-proxy-server-with-portal-for-arcgis.htm.  ... View more