Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Disappointed with some of the changes to NPM 10.1.1

I hate to be negative, but I've really got to provide some input. i am very unhapy with some of the changes with orion NPM10.1.1.

I find the web interface very much slower and more difficult to use that the System Manager application. I use the System Manager application for many purposes. I did not like it when SolarWinds made it so that you have to use the web interface if you want to modify the name of a node or IP address (or anything that was available through Node Details). But, that was months ago, version 9.5 I think.

Now, SolarWinds has further limited the usefull ness of System Manager. Prior to 10.1.1, you could review the Orion NetPerfMon Event Log by clicking on an icon on the tool bar in System Manager; now you can't. Previously, you could right-click a node or interface and instruct it to Poll Now; now you can't. Previously, you could right-click a node and change which Resources are monitored; now you can't. Previously, you used to be able to add a node to NPM from the System Manage; now you can't.

I'm sure there are many other functions that previously were enabled, but have been disabled in version 10.1.1, but it seems as if the idea has been conveyed. I'm rather upset about this all because these changes have made my job harder! I don't understand why SolarWinds is so intent on disabling the OS-running version of an interface to Orion, and forcing people to use a web based front end. It was good when the System Manager was fully enabled with all the features you could want, but allow people to publish the information via Web interface. I'm all for allowing web interaction! But the idea of disabling features on the native OS-running application in order to force users to use the web interface is just ludicrous. Web interfaces are a good idea if they are AN option, not the ONLY option!

PLEASE PLEASE PLEASE re-empower the System manager GUI -- don't force us to use the web interface.

Eric Bradford

Find more posts tagged with

Accepted answers

All comments

bshopp

Eric,

Let me not beat around the bush, System Manager is not coming back. In fact, it is the opposite, it will ultimately go away. See this post I did for more background, but I will also take some time to explain here.

So why are we doing this? Because I am an evil evil man? No, I do not get pleasure of seeing you angry about NPM 10.1. There are multiple reasons why we are doing this, some of which are changes you are seeing, some are not as they are back end architectural changes. I will walk through one of the reasons here.

In NPM 10.1 we introduced tons of great features as highlighted here. Two of which are VMWare ESX integration and Cisco UCS Monitoring. Both of these features require us to poll not via SNMP, but through their API. So we need a user to enter and manage these credentials both within add/edit node, but also if you look under VMWare settings there are some additional credentials management there.

So you say, fine I get that, add them to both the web and System Manager. To do this is not a trivial amount of work and if you add this up vs. some other features which fit into this bucket then this ultimately means we can do less new features and bug fixes for you, the end users.

Our strategy is to move more and more to the web to allow for easier remote management. Instead of having to RDP to the server, just connect to the web console. When we started these changes in 10.0, some users had a very similar reaction as you did. We asked them what was hard to use and feedback was provided and since then we have made enhancements to both performance on this page as well as user flow to perform actions. Everything you could do in System Manager, you can do in Node Management in the web. Tell us specifically what is hard to do/find or where it is slow and we can work on enhancing that. Since we have started making these enhancements, many of the users who originally were angry have been converted and like the web much better.

The old System Manager had many limitations, one of the largest was you could not do multi-edit, which you can do in the web. This is just one example. The web is much more feature rich than System Manager ever was.

We also tweaked other areas in the web. You can do poll now from web node management or directly from the node details page. This is also the same case for List Resources.

We have also introduced a new page called Message Center in 10.1 which combines the Orion Events, Alerts, Syslog and Traps into a single page which allows you do filter, sort etc.

I know my answer does not ultimately make you happy, but hopefully this sheds some insight into the method behind the madness. I am more than happy to chat offline with you further about this via email or the phone and please please keep any and all feedback coming at us. You may not always like the answer I give, but trust me, we are listening and taking it to heart.

netwiz2007

I actually have no issue with moving to a web based interface, however if this is the direction (and it clearly is) then Solarwinds MUST do something about the performance of the WEB UI. We are running NPM 10.1 on a very high end server 18 cores and 16gig of ram and the web ui is still slow.

rsprim

Netwiz,

It's very dissappointing to see your post considering we've been complaining about web performance since 9.5 and were assured that performance increases were built into 10.

I haven't completed the upgrade to 10.1.1 yet but I get the impression based on your response that once I do I'll continue getting hammered daily regarding the WEB UI performance. I am also conditioned at this point to hear the standard response from SolarWinds of "It will be fixed in the next release". That is unfortunately what we have to tell our customers too and they are tired of hearing it.

While I appreciate all the new features that Solarwinds puts out and the frequency with which they do so, none of that really matters to my operational folks if it takes them too long to access the information because the interface is slow. Our Ops folks call the system "SlowlarWinds" which really bugs me but until SolarWinds speeds up the interface, there is not much I can do. Hopefully, SolarWinds can and will make the UI faster so I quit getting beat up.

bshopp

One tip on web performance to try out that I have seen that helps folks.

If you have real time anti-virus scanning enabled on your Orion server, be sure to exclude the inetpub directory from these scans.

On the web performance side, I hear you loud and clear. We did some increases in 10.0 as well as spent some development cycles in 10.1 looking at the web console and making changes to where we can be more efficient. However, point in taken, there is still work to be done.

byrona

I am curious, are the people that are experiencing the slowness of the web interface using the web server located on the core NPM system or are they using the additional web server component?

I personally found the web performance to be a bit slow (not unusably slow) on the core NPM system but we use the additional web server and with that our web performance has been pretty good. Overall I still find Orion to be on the slow side as far as web UI's are concerned but with the additional web server it's definitely functional considering how many people we have using it.

NCM on the other hand is slow to the point of almost being unusable for us.

As far as the direction to be completely web based is concerned, I fully support that direction as it provides much better multi-user support and is consistent with the direction most modern multi-user systems are going. So long as measures are taken to keep the interface responsive and user friendly.

LGentile

Me and the members of our support team have gone back and forth on this issue. I'll summarize our love/hate relationship, but know that I agree with SW's direction of moving all to the Web console in terms of development.

- Having 30+ people log into multiple polling servers was horrible. The Web console fixed this. One login, one portal, one to rule all your pollers.

- It is actually easier to update multiple nodes for custom properties, etc now using the Web console when it is performing 100% (more on this later).

- Adding pollers to a mass amount of devices is also much easier now, as is adding nodes in general because you get a 'wizard' of sorts. Before, people could 'forgot' to launch CPE and add custom properties - a battle i fought for a LONG time.

- The customizable views within Manage Nodes has made my life (as the main SW admin) a ton easier. Truth be told, I RARELY use System Manager anymore, and I was probably one of the loudest screamers that this functionality went away on my team

- Our performance issues seemed to revolve around SQL. We do have SQL running on a enterprise platform shared with other databases, however our SQL team found that our DB was highly fragmented. They created a job to clean this up when fragmentation reached 30% and I can tell you this made a drastic increase in performance. Now the only real issues I see are related to javascript errors, but retrying the command usually works the second time so I haven't bothered to follow up with it. I am now able to add a hardware poller set to 100+ devices at a time and it works fine (before SQL tuning, i could barely get an update pushed to 20 nodes at a time).

I don't think you stated your configuration, but we are all virtual except for the SQL server and use an additional Web server. That web server basically sits there twirling its thumbs all day long - basically NO CPU usage, so all our issues were SQL related. The pollers are naturally very busy but this is mainly why we run a separate Web server.

ebradford

As many people here, i have been using SolarWinds for several years, at two different companies. I think I started with Orion 8.5. I have been using the web interface and the System Manager. Both have their uses.

I want to clarify my complaint -- it's not that I'm looking for SW to increase the productivity of SM; it's not "if only the web GUI was faster..." -- it is that SW removed functionality from SM. Also, I am the only Orion administrator at my company. However, I allow all the other IT peeps to access Orion via the web interface. I'd never let them into my server.

bshopp -- I can understand that adding UCS and VM polling would make it that you can't add a VM or UCS node using the old method in SM. But that doesn't logically mean that the SM method must be removed. You should still allow "regular" nodes to be entered in the old manner.

I also can't understand how the changes you refered to would necessitate not allowing people to view the NetPerfMon Event Log. Are you planning on removing the Syslog viewer too? Why must we use the web GUI to view the Even Log? Why can't you just allow us to use an OS front end for it? Also, SW spent some time an programming skills to create a new OS front end for SW Tools. Nice. I like it. If SW was so pro-web, why do that? Shouldn't it have been better to create a Web Workplace Studio? The functionality you newly created for SW Tools, should be retained for Orion.

Speed of the web interface does happen to be an issue. But previously, it was just an issue for other IT people, not me. I could live with that. Now speed is an issue for me too. I can't live with that. Call me self-concerned...

When adding a node, I used to be able to click "add node", pop in the IP and the SNMP string, and be done in less than 10 seconds (not including selecting resources). Now it takes 20 seconds for just the first screen (to enter IP etc) to show up in the Web Gui. Plus all the other screens. that needlessly must be navigated and waited for before craeting a node. Adding 6 nodes I could do in a minute, now takes me ten.

If SW really is going to get rid of SM (I advise against it), then I respectfuly submit these suggestions:

1. On the add-node screens, have the "next" button at the top of the screen as well as the bottom.

2. Have a "Finish" button on each page, so if we don't want to select resources or if the node isn't VM (etc), we can just finish the add without having to go through unnecessary screens. Or, have buttons that reflect how you want to proceed: Finish, Resources; UCS; VMWare.

3. Create some more Orion OS fronts ends for specific uses. These would be similar to Report Writer, Database Manager, Test Fire Alerts, Report Writer, Custom Properties Editor, etc. For goodness sake don't get rid of these! The new applets could be: NetPerf Event Log Viewer; Node Adder; Instant Node Poller.

4. Make Orion Web run really fast

netlogix

Oh, like ebradford said, the number of screens can be a little annoying, or beneficial. The ability to have the "Add Node" wizard (?) be a little more customizable as far as how many screens you go through would be awesome. I like the one that show me all the Custom Properties for me to fill out, but some one else may not want it. I was able to remove one page, so now I only have 3 pages per node.

ebradford, if you are going to add 10 nodes, you might want to use the network sonar. You can create a one time network sonar and give it a list of IPs.

if I were to add 1 - 4 nodes, I would go through the standard web way and that is currently 3 - 12 pages. But 5 - 10 is bad, 15-30 pages, but sometimes that is easier than trying to get sonar to actually select the right resources. i.e. windows 2008 has umpteen NICs, but only one is real.

I tried opening multiple tabs to add multiple server, but that turned out really ugly.

If the Add node was just one page, but had button to configure each other thing, then that would be awesome!

bshopp

This is good feedback keep it coming. To the earlier point on why remove existing functionality I already discussed one reason, but another reason is if you noticed in 10.1 we introduced three new services with the word collector in them. We are starting to introduce a next gen poller into the product which will be multi-threaded etc. We ported over some of the things the standard poller collects to this new collector in 10.1 and plan to move the rest over going forward. So we would have to update System Manager existing functionality to talk and work with this new collector. Which again means less new features and bug fixes.

netlogix

Oh, and like NCM's manager program, the page could keep re-loading empty and be ready to add another node after each time you click add.

FormerMember

...our SQL team found that our DB was highly fragmented. They created a job to clean this up when fragmentation reached 30% and I can tell you this made a drastic increase in performance.

LGentile, would you be willing to share the job?

If the Add node was just one page, but had button to configure each other thing, then that would be awesome!

+1 on this one.

Rhodizzle

Thought I'd post this since I found a good resolution. I recently upgraded my 10.0 installation to 10.1 and also took the opportunity to migrate the database to a centralized SQL server (We have to do the DoD STIGs so the less SQL STIGs I have to do, the better). Anyways, the performance on the WebUI was terrible on all pages. I started researching defragging the database and found the below script, which has worked EXTREMELY WELL. After running this on my NetPerfMon database the WebUI has returned to the performance I had with 10.0 and a local DB server.

/* Originally created by Microsoft */
/* Error corrected by Pinal Dave (http://www.SQLAuthority.com) */
-- Specify your Database Name
USE NetPerfMon;
GO
-- Declare variables
SET NOCOUNT ON;
DECLARE @tablename VARCHAR(128);
DECLARE @execstr VARCHAR(255);
DECLARE @objectid INT;
DECLARE @indexid INT;
DECLARE @frag decimal;
DECLARE @maxfrag decimal;
-- Decide on the maximum fragmentation to allow for.
SELECT @maxfrag = 15.0;
-- Declare a cursor.
DECLARE tables CURSOR FOR
SELECT CAST(TABLE_SCHEMA AS VARCHAR(100))
+'.'+CAST(TABLE_NAME AS VARCHAR(100))
AS Table_Name
FROM INFORMATION_SCHEMA.TABLES
WHERE TABLE_TYPE = 'BASE TABLE';
-- Create the table.
CREATE TABLE #fraglist (
ObjectName CHAR(255),
ObjectId INT,
IndexName CHAR(255),
IndexId INT,
Lvl INT,
CountPages INT,
CountRows INT,
MinRecSize INT,
MaxRecSize INT,
AvgRecSize INT,
ForRecCount INT,
Extents INT,
ExtentSwitches INT,
AvgFreeBytes INT,
AvgPageDensity INT,
ScanDensity decimal,
BestCount INT,
ActualCount INT,
LogicalFrag decimal,
ExtentFrag decimal);
-- Open the cursor.
OPEN tables;
-- Loop through all the tables in the database.
FETCH NEXT
FROM tables
INTO @tablename;
@FETCH_STATUS = 0
BEGIN;
-- Do the showcontig of all indexes of the table
INSERT INTO #fraglist
EXEC ('DBCC SHOWCONTIG (''' + @tablename + ''')
WITH FAST, TABLERESULTS, ALL_INDEXES, NO_INFOMSGS');
FETCH NEXT
FROM tables
INTO @tablename;
END;
-- Close and deallocate the cursor.
CLOSE tables;
DEALLOCATE tables;
-- Declare the cursor for the list of indexes to be defragged.
DECLARE indexes CURSOR FOR
SELECT ObjectName, ObjectId, IndexId, LogicalFrag
FROM #fraglist
WHERE LogicalFrag >= @maxfrag
AND INDEXPROPERTY (ObjectId, IndexName, 'IndexDepth') > 0;
-- Open the cursor.
OPEN indexes;
-- Loop through the indexes.
FETCH NEXT
FROM indexes
INTO @tablename, @objectid, @indexid, @frag;
@FETCH_STATUS = 0
BEGIN;
PRINT 'Executing DBCC INDEXDEFRAG (0, ' + RTRIM(@tablename) + ',
' + RTRIM(@indexid) + ') - fragmentation currently '
+ RTRIM(CONVERT(VARCHAR(15),@frag)) + '%';
SELECT @execstr = 'DBCC INDEXDEFRAG (0, ' + RTRIM(@objectid) + ',
' + RTRIM(@indexid) + ')';
EXEC (@execstr);
FETCH NEXT
FROM indexes
INTO @tablename, @objectid, @indexid, @frag;
END;
-- Close and deallocate the cursor.
CLOSE indexes;
DEALLOCATE indexes;
-- Delete the temporary table.
DROP TABLE #fraglist;
GO

mr.e

I do not mind having System Manager within the Windows interface. although I personally favor the Web UI instead and look forward to one day using the Orion web console exclusively. I just see many benefits for using a Web-based app rather than a Windows-based one, even if at times we deal (and/or struggle) with performance issues on the Orion web console. Unfortunately, due to limtiation in man-hours at SolarWinds, it is probably very hard for the SolarWinds programmers to keep developing and maintaining two flavors on the same Orion NPM application (a Windows-based and a Web-based). This is a though balancing act.

Now, there is a related point that I would like to make. I do appreciate the large number of free tools that SolarWinds develops. These probably add value for many, and are nice ways to attract new customers. However, if there is a resource issue for the SolarWinds programmers, perhaps it would be best for them to reduce somewhat the number of free tools. This way, they could spend more cycles working on programming for the paid tools instead.

Perhaps, if the SolarWinds developers do this, maybe they can continue to maintain both Windows and Web-based platforms for the Orion tool suite for a bit longer. That is, until the Web version has equal or better tools than the Windows-based one. For example, I very much look forward for the Report Writer function to be added to the Orion web console. This would keep us from having to grant local rights to persons outside my team.

acherman

I'm with Eric on this one - I miss the functionality of the System Manager. My biggest issue is also with slow web UI performance. I liked having the Event Viewer up in the SM - it's updates/refreshes appeared to be more like real-time updates, rather than pressing refresh or waiting a minute or two for web pages refreshes. Disappointing. I will attempt to defrag our DB as well to see if any performance returns.

One thing to note, I believe the release notes for NTA 3.7 specify performance increases, but I don't see where.

Aaron

FormerMember

well done mate, love this sort of stuff....

is there anything we should be aware of when running this? did you run it while orion was up and running? i have 5 additional loaded pollers, so do not want to cause a catastrophic failure...

could the solarwinds staff review this and acknowldge it as a supported practise?

how long did it take to run? how big was your DB? do you run in full recovery or simple mode for SQL?

cheers

dan

coda6

I know this is an old and dead thread but I found that when using the NPM web interface in IE it's incredibly slow. Almost to the point of click on a node and go get a cup of coffee slow.

Then I tried using it in Chrome. Night and Day difference. Chrome loads our Network Summary page in under 5 seconds. IE9 was just under 15 seconds.

For what it's worth.

Ken

bmvernil

Has anyone resolved this issue? Same problem here moved the database server and the application server to new updated hardware and the performance stinks.