I have two sets of installations recently upgraded to 8.2.0. In one set, when I click "Create New Dashboard", it presents the choice of "Classic Dashboards" and "Dashboard Studio", i.e., as expected in 8.2. But the other set presents no such choices. What can I do to find out why?
The following table tries to summarize differences that I can think of.
A1 | A2 | B1 | B2 | |
Dashboard Studio | functional | - | dysfunctional | - |
OS | MacOS | - | Linux | - |
Firewalled | no | - | yes | - |
Port | 8080 | - | 443 | - |
SSL | no | - | yes | - |
Upgrade path | 8.1.2 -> 8.2.0 | 6.5.1 -> 8.2.0 | 8.1.2 -> 8.2.0 | - |
In both sets, when I go to "Manage Apps", then search for Dashboard Studio, it shows up and I can "launch" it.
Actually yes.
In your system/default/web.conf you should have these conf:
# First party apps:
splunk_dashboard_app_name = splunk-dashboard-studio
# enable/disable splunk dashboard app feature
enable_splunk_dashboard_app_feature = true
Please check if you are overwriting them somewhere (e.g. system/local)
I had a copy of the old web.conf in system/local so the new configurations were not applied.
We were able to resolve this issue on the box that had it by modifying the web.conf file, this was on Rocky Linux.
Edit: /opt/splunk/etc/system/local/web.conf
changed:
# First party apps:
splunk_dashboard_app_name = splunk-dashboard-app
to:
# First party apps:
# splunk_dashboard_app_name = splunk-dashboard-app
splunk_dashboard_app_name = splunk-dashboard-studio
enable_splunk_dashboard_app_feature = true
Hi,
same problem here.
Upgrade path: 8.1.3 -> 8.2.2.1
Is there any news?
Thanks!!
Any updates on this. Having the same issue on CentOS based system.
Actually yes.
In your system/default/web.conf you should have these conf:
# First party apps:
splunk_dashboard_app_name = splunk-dashboard-studio
# enable/disable splunk dashboard app feature
enable_splunk_dashboard_app_feature = true
Please check if you are overwriting them somewhere (e.g. system/local)
I had a copy of the old web.conf in system/local so the new configurations were not applied.
@davidemagni wrote:I had a copy of the old web.conf in system/local so the new configurations were not applied.
Thank you! I copied a few lines too many under [settings] from default/web.conf when setting up custom cert. As such, subsequent changes in defaults that happened to be copied over are not reflected. Again, maybe some specific pointers existed in 8.2.0 upgrade instructions, but if they did, I clearly did not take notice.
Simply removing splunk_dashboard_app_name = splunk-dashboard-app made studio available to me. Now I am stripping my local/web.conf to the minimum required.
I guess there's still been no progress on this? Just upgraded both of our development environments from 8.1.5 to 8.2.2 and seem to be missing the option in both. The only way to get to the new dashboard is through directly clicking "Launch App" on the "Manage Apps" page.
Hi there, did you have any luck figuring this out? I'm in the same position, Clone to Dashboard Studio option is also available but it does load the dashboard studio, when I leave though it cannot be returned to.
Mine (problematic ones as well as A3 the functional one) are all CentOS 7.9. So we know the glitch can happen on different distributions. Anything in upgrade path that is in common with the mine?
I can't recall exact versions but the upgrade path was 8.1.x > 8.2.0 > 8.2.1. It might have been 8.1.3.
@2422z9s8 Unfortunately I haven't made any progress. But now at least we know this can happen in more than one instance. What is your OS?
Hi @yuanliu are you able to open a support case for this issue?
Hi. As an experiment I upgraded from 8.1.3 to 8.2.1 and we use a third party for certs and things DID work.
I strongly recommend a Splunk support case be opened! Splunk support can help turn on extra debugging to track this issue down.
**bleep**! I'm running 8.2.1 on Ubuntu 20.04. Didn't show up on 8.2.0 either.
After considering the diagnostic matrix, I suspect that SSL configuration may be critical. To test this, I created a new Linux server behind the same firewall, and tested default installation, i.e., plain text HTTP on port 8000. Lo and behold, Dashboard Studio option shows up.
So, the question now is: how to diagnose my SSL configuration to fix this? (I use my custom certificate, not Splunk's default.)
@yuanliu wrote:So, the question now is: how to diagnose my SSL configuration to fix this? (I use my custom certificate, not Splunk's default.)
To confirm, I enabled SSL on the newest installation using out-of-the-box self-signed cert. Dashboard Studio shows just fine.
@yuanliu wrote:
@yuanliu wrote:So, the question now is: how to diagnose my SSL configuration to fix this? (I use my custom certificate, not Splunk's default.)
To confirm, I enabled SSL on the newest installation using out-of-the-box self-signed cert. Dashboard Studio shows just fine.
Unfortunately, the hypothesis about SSL configuration is still incorrect. In the newest installation (I'll call it "A3"), I reconfigured serverCert, privKeyPath in web.conf, and, serverCert, caCertFile in server.conf just as I did in B1 and B2, so both Web GUI and REST API/admin server use a custom cert. Dashboard Studio is still available in A3. The diagnostic matrix is now
A1 | A2 | A3 | B1 | B2 | |
Dashboard Studio | functional | - | - | dysfunctional | - |
OS | MacOS | - | Linux | Linux | - |
Firewalled | no | - | yes | yes | - |
Port | 8080 | - | 443 | 443 | - |
SSL | no | - | yes | yes | - |
Upgrade path | 8.1.2 -> 8.2.0 | 6.5.1 -> 8.2.0 | fresh 8.2.0 | 8.1.2 -> 8.2.0 | - |
The only recognizable difference is upgrade path. This points to possible glitch during the upgrade for those two instances. (B2 is just a test instance that is configured nearly identically to B1 except the actual cert used and ingested data.) Even this doesn't quite explain much, as A1 is also upgraded from 8.1.2. (If anything, the upgrade from 6.5 on A2 should be more problematic.)
Has anyone been able to resolve this issue? I have the latest version of Splunk 8.2.1. I currently have a similar setup with custom SSL Certificates and using port 8443 to access Splunk via the web.
An examination of etc/system/local/ on A3 and B1 gave me no meaningful explanation why one would work while the other not. They are both very vanilla. splunkd.log, web_service.log also do not show any error or warning when invoking dashboard. What else can I check?
As I tinker, the plot gets thicker. I copied sample_app and built a blank new app. From this app, if I navigate to "Default Views -> dashboard", I enter a vanilla Dashboard Studio dashboard.
I can edit this dashboard in Dashboard Studio (no choice), I can save it. But the saved dashboard cannot be viewed anywhere, and is not listed in "Default Views -> Dashboards". However, it is listed in that "Default Views" drop-down; when I click that item, it says 404 page not found. (I checked the view path and dashboard xml definitely exist.)
If I use "Default Views -> dashboard" again, it presents another blank Dashboard Studio dashboard.
In addition to "Default Views -> dashboard", the "Clone to Dashboard Studio" option is also available for any Simple XML dashboard. However, if I use this, the result is the same: 404 page not found even though an XML file is created right in the same directory!
2021-06-12 17:12:56,957 INFO [60c4eb18de7fa2e47b3850] cached:163 - /opt/splunk/etc/apps/splunk_essentials_8_2/appserver/static/setup.json
2021-06-12 17:12:57,122 INFO [60c4eb18de7fa2e47b3850] error:321 - Masking the original 404 message: 'Nothing matches the given URI' with 'Page not found!' for security reasons