Chat transcript from Office Hours for Spectrum - May 2015 which took place on May 28 2015.
Melanie Giuliani to Everyone: Good morning everyone - welcome to Office Hours for Spectrum! We'll be getting started in about a minute.
Ricardo to Everyone: Just confirming: It seems this meeting doesn't have audio, does it?
Stuart Weenig to Everyone: @Ricardo - No, the office hours never have audio. Chat only.
Melanie Giuliani to Everyone: Hi @Ricardo - yes, you're correct. This is a Webex chat only event.
Melanie Giuliani to Everyone: Alright, let's get started! Please start typing your questions into the chat box.
Stuart Weenig to Everyone: I'm looking for information on best practices around device management. Specifically: https://communities.ca.com/thread/241731702
Melissa Potvin to Everyone: @Stuart, the team is reviewing the thread, please hold just a moment
Stuart Weenig to Everyone: @Melissa - np
Stuart Weenig to Everyone: I'm looking for functionality like i have in NV: NV maintains a discovery profile, which is just a list of IP addresses. If there is a device in NV that isn't in the discovery profile, it marks it as out of scope. It's really easy to find those devices, figure out which poller (landscape) it should be in and either add it to the scopes list (dp) or delete the device.
Joseph Ackley to Everyone: @Stuart the discovery results can be configured to export to csv file
Stuart Weenig to Everyone: @Joseph, can i export all of them via command line with one script on the mls or would i have to do it on each SS?
Neville Styles to Everyone: I asked a question in the community but had limited (although useful) answers. We need to monitor a system that can only send exceptions via email, no syslog, traps etc as it's a cloud-based solution. How would you go about generating events in Spectrum from an email?
Deb to Everyone: With the Spectrum UIM/Nimsoft integration has there been any thought to add more specific cause codes for UIM/Nimsoft alarms coming into Spectrum?
Sahoo to Everyone: @Deb - What specific cause code we are referring here ?
Jason Meader to Everyone: @Neville - two options come to mind -- one is to have the email log to a file, and then use log file monitoring...the second is to convert the email to xml and then use the southbound gateway integration
Deb to Everyone: @Sahoo - Today, there is about 5 different cause codes, one for each severity level in Spectrum, and if there could be more specific cause codes, one for each probe even would be helpful to be able to run some specific policies/rules against in SANM
Sahoo to Everyone: @Deb - Please list down the specific scenario and cause codes, and what is the expectations , then we can take forward
Nagesh Jaiswal to Everyone: @Deb: We are thinking towards the direction with respect to Spectrum-UIM integration. You will hear soon on this from Product Management team.
Nghia Van to Everyone: Is there a plan to allow clustering of NMS?
Nagesh Jaiswal to Everyone: @Stuart: We do not have any inbuilt Spectrum functionality to know from which discovery configuration a model was modeled from
Neville Styles to Everyone: Thanks @Jason. The tricky bit is how to read an email, writing to a log or converting to XML is relatively straightforward after that. Is there some sort of mailbox monitoring in the CA suite that could achieve that?
Stuart Weenig to Everyone: @Nagesh Ok, that answers that. How would I invoke the modeling gateway to export all the DP's across all landscapes on the MLS server?
Peter to Everyone: What Operational Metrics Reporting are you planning?
Nagesh Jaiswal to Everyone: @Stuart: a small perl utility, which will export all discovery configurations (only),
Todd Kornely to Everyone: @Nghia, can you clarify what you mean by clustering of NMS?
Ashok to Everyone: @Stuart, I have a perl script which will export all DP's across all SS within a DSS
Stuart Weenig to Everyone: @Nagesh: Yeah, as you can see, I've already written the perl to parse through it. I'm starving for documentation around how to get the modeling gateway to give just what I want.
Stuart Weenig to Everyone: @Ashok, please post it!
Ashok to Everyone: @Stuart, that should help your situation
Stuart Weenig to Everyone: @Ashok, yes, I think that would help.
Nagesh Jaiswal to Everyone: @Stuart: We will respond to your community post
Nghia Van to Everyone: @Todd - The primary nms system that is running data_engine, we would like a way to have more than one instance of this. further more, rather than running a secondary HA have 2 or more instance of the primary server running in tandem at all times
Ashok to Everyone: @Stuart, sure we will respond to your community post
Stuart Weenig to Everyone: @Ashok, ty
Deb to Everyone: @Sahoo: for example 6330001 is for Minor UIM alarms, 6330002 Major, 6330003 Critical... If we could have 6330010 for CDM probe, 6330011 for process probe, etc... then we could apply rules/policies agains those cause codes within SANM - just to isolate to the sev level doesn't allow for more flexibility with the alarms coming in. Would even like to get more specific with like CDM probe having cause codes for Disk, CPU,Memory, etc...
Todd Kornely to Everyone: @Nghia, I am not aware of any plans for clustering like this.
Nghia Van to Everyone: @Todd - short of the ability to cluster, is the only way to also fail over wasp and such things that depend on data_engine only accomplished via some custom srcipting?
Sahoo to Everyone: @Deb we should able to do it, will have internal discussion and get back to you
Stuart Weenig to Everyone: Is there a way to easily identify duplicate devices? I mean devices that were on one landscape but have been accidentally added to another?
julio to Everyone: Talking about UIM and Spectrum integration.. I'm wondering why Spectrum is allowing the UIM integration to change the model name if the model already exist in Spectrum.. for example.. if I had a model for a server in Spectrum called myserver.org.com an the same server in UIM is labeled myserver then Spectrum changes the model name to myserver. Is this by design? I would rather have an option in the integration to be turned on if this is wanted
Manish Upadhyay to Everyone: @Nghia Van, for every database instance there will be only one data_engine, if you want to have mulitiple data_engine you can have multiple database instance however multiple data_engine can be in one domain..
Nagesh Jaiswal to Everyone: @Peter: We will look into it and get back with an update later.
Ashok to Everyone: @Stuart, one way is locater search
Stuart Weenig to Everyone: Ashok, How?
Todd Kornely to Everyone: @Nghia, are you referring to Spectrum? I am confused about your data_engine references.
Ashok to Everyone: @Stuart, from the locater tab, search for a device with IP , which will retrieve all devices from all landscapes
Peter to Everyone: Are there plans to support DMVPN (and Live Links)?
Stuart Weenig to Everyone: @Ashok, ah, and then export, use excel formulas to find the duplicates?
Nghia Van to Everyone: @Todd - we are currently running our database in a sql 2012 cluster and pointing to the cluster alias name so for all intents and purposes, UIM sees it as one database server, also we only have one domain. The end goal is the ability to easily scale our solution as we are a multi-tenant MSP (also some of our customers are multi-tennant MSP themselves)
Ashok to Everyone: @Stuart, correct
Stuart Weenig to Everyone: @Ashok, I know i can search for a single device and find out if there are duplicates. I'm looking for a way of finding all devices that I don't know are currently duplicates.
Nagesh Jaiswal to Everyone: @Peter: This is a new enhancement request. Request you to have and idea open on Community wall
Melanie Giuliani to Everyone: @Peter - here is a link to the community where you can create an idea on that: https://communities.ca.com/community/ca-infrastructure-management
Jason Meader to Everyone: @Julio - yes, it does currently function this way in that the Nimsoft name wins. I do believe we are looking into a future release that will allow options for the naming
Ashok to Everyone: @Stuart, we currently do not have any such method to find duplicate models, that would be an enhancement request
Peter to Everyone: @Melanie & Nagesh - Have already. Thanks
julio to Everyone: @Jason - That will be very useful.. Thanks
Stuart Weenig to Everyone: @Ashok: done - https://communities.ca.com/ideas/235722602
Ashok to Everyone: @Stuart, thank you
Nagesh Jaiswal to Everyone: @Peter: Thanks. If you can provide the link I will get it reviewed.
Melissa Potvin to Everyone: OK, we are almost at the 1/2 hour mark. Great first half. Keep the questions coming!
Peter to Everyone: @Nagesh - hmmm was a while ago
Nagesh Jaiswal to Everyone: @Peter: No Problem. I will search and respond to the community post. Thanks.
Catalin Farcasanu to Everyone: Is there any timeframe available for the release of 10 version?
Peter to Everyone: @Nagesh - try last year (Dec +/-)
Stuart Weenig to Everyone: Funny juxtaposition of those two comments.
Stuart Weenig to Everyone: @Catalin - Last i heard was sometime near June. Completely unofficial though. They can't release that stuff until they actually release it.
Nagesh Jaiswal to Everyone: @Catalin: It will be available in 2nd half CY15
Melissa Potvin to Everyone: If you are new to office hours or just joined us, the support team is standing by to answer your questions. Use the chat window, type your question and hit send.
Catalin Farcasanu to Everyone: June is around the corner. Thanks @Nagesh!
Stuart Weenig to Everyone: @Catalin: which basically means the next few days, or June.
Nagesh Jaiswal to Everyone: @Peter: Thanks. I will look into it.
Stuart Weenig to Everyone: oh. second half, nevermind. guess it got pushed back.
Stuart Weenig to Everyone: so that could be december.
Melanie Giuliani to Everyone: For those interested, here is a link to the replay & content for our recent webcast on Spectrum 10.0 https://communities.ca.com/community/ca-infrastructure-management/blog/2015/04/22/webcast-spectrum-100
Stuart Weenig to Everyone: so peter's comment was actually right
Stuart Weenig to Everyone: but this year.
Peter to Everyone:
Karan Nirola to Everyone: @Stuart - we are on time :-)
Catalin Farcasanu to Everyone: Who can tell me if there are any improvements on REX integration with the new version. And maybe what's the latest of the REX version that is certified to work with SPECTRUM.
Stuart Weenig to Everyone: @Karan - On time is relative. The first time i heard anything more definitive than 'this year' was March 2015. Then April 2015. Then May 2015, then June 2015. Now second half of this year, which is July-December.
Nghia Van to Everyone: also when will there be support to cluster unified reports in the uim portal since the portal itself can be clustered but the unified reports which is accessed via the ump portal cannot
Nagesh Jaiswal to Everyone: @Catalin: REX 11.x is the support coming with Spectrum 10.0
Jason Meader to Everyone: @Nghia - that would be a good question to post to the communities as it's in regards to the UIM product and not Spectrum
Nghia Van to Everyone: im sorry is this office hours only for spectrum?
Melissa Potvin to Everyone: @Nghia, yes this office hours event is for CA Spectrum.
Catalin Farcasanu to Everyone: @Nagesh: any improvements, other than support for the new version? Maybe some capabilities extension?
Melissa Potvin to Everyone: @Nghia, post your question on UIM here: https://communities.ca.com/community/ca-infrastructure-management and be sure to tag it to UIM category.
Hans Driessens to Everyone: Hi All, Is there a way for spectrum to remember the location of off-page references if a device/model is removed/added. A lot of data is kept (e.g old configs) but off page references and the locations not.
Hans Driessens to Everyone: with location I mean place in the topology view
Nagesh Jaiswal to Everyone: @Catalin: For now it is just certification effort. If you have any new features in mind then would request you to open idea on Community wall. We will look into it.
Joe Poutre to Everyone: Does anyone have a good way to map back SANM policies to Global Collections and/or Models? I have a time trying to trace back the reason why some group received an unexpected email, and more than once deleting a G.C. ended up causing a SANM policy to go crazy because it now applied globally. Another way to put it, we need a way to see what SANM policies apply to a given G.C. or a given Model; we use Model-specific SANM policies for SPM tests.
Stuart Weenig to Everyone: @Hans, the same applies when moving the model the off page reference refers to. If you move the model from one container to another, the off page references to that model reset everywhere.
Todd Kornely to Everyone: @Hans - why would you want to keep OPR location? For when you remodel the device?
Stuart Weenig to Everyone: @Joe, can't you do that just by opening the SANM policy and going to the Collections tab?
Joe Poutre to Everyone: @Stuart - Because we have many SANM Policies on three different Applications, due to various escalation requirements and different email format designs. I have to search through all of them each time I want to clean up unused G.C.s.
Hans Driessens to Everyone: @Stuart - 2 reasons, 1) sometimes strange behaviour of the model (not so often as it used to be) and 2) sometimes I have to move a newly discovered provider_cloud model t
Stuart Weenig to Everyone: @Joe, ah, understood. I have 1.
Hans Driessens to Everyone: (oops) to a new location as I don't want them in the root of the universe.
Joe Poutre to Everyone: @Stuart - And for the SPM tests, which are mapped by Model name, there's no easy way to check that each is in an SANM policy, nor to find out to which policy a given SPM test is mapped. We have to track it manually using a (gak!) spreadsheet.
Stuart Weenig to Everyone: @Joe, wow. Sympathy for reals.
Todd Kornely to Everyone: @Hans, I do not think it is possible to persist this. Sorry
Hans Driessens to Everyone: @Todd - Is there any way to set a different location besides the universe root for newly discovered Provider Clouds using the EVPN discovery?
Melanie Giuliani to Everyone: This is the 15 minute warning! Please get your last minute questions and follow-up responses in now!
Todd Kornely to Everyone: @Joe, if you are looking for a way to quickly determine what policy caused a notification, I would use NotificationData for this( have each policy have unique notification data, so you know what policy caused the notification )
Joe Poutre to Everyone: @Todd - that will help one way, but I also need the other way - what SANM policies apply to a given G.C? - without having to check all of them or keep a spreadsheet.
Todd Kornely to Everyone: @Hans, have you tried setting Destination container in Adisc modelling?
Stuart Weenig to Everyone: @ Todd – Adisc modeling?
Todd Kornely to Everyone: @Joe, I don't know of a way to do this curently. Would suggest post on idea wall for enhancement
Todd Kornely to Everyone: @Stuart, Modeling Options>Destination Container
Hans Driessens to Everyone: @Todd provider cloud models are not affected by that, the EVNP manager has its own behaviour
Stuart Weenig to Everyone: @Todd, ok. was confused by Adisc.
Stuart Weenig to Everyone: @Todd, we did that. We set the target container to be the 'New Devices' container, but the clouds still all appear at the root of the universe. I figured it was by design.
Todd Kornely to Everyone: @Hans, not sure if per design or not, but seems you certainly would like different behavior
Todd Kornely to Everyone: @Hans, but thinking more about this, since the models are created much later in the adisc process, the destination container probably doesn't apply. If this is important to you, I suggest posting on the idea wall
Melanie Giuliani to Everyone: Alright guys, that wraps up today's session. Thanks for your questions! Today's transcript will be posted on the IM community.
Hans Driessens to Everyone: @ Todd - will do
Stuart Weenig to Everyone: @Hans, if you put the idea, mention me in the comments (@sweenig) and I'll vote it up.
Joe Poutre to Everyone: FYI: I've posted on the Communities about the SANM policy mapping issue.
Stuart Weenig to Everyone: @Joe - link?
Joe Poutre to Everyone: https://communities.ca.com/thread/241732328