Chat transcript from CA UIM Office Hours: A Live Online Chat [July 2015] which took place on July 7, 2015.
Aditya to Everyone: Hi All, Good Morning
Melissa Potvin to Everyone: Hello Everyone! Thank you for joining office hours today for UIM. We will get started at the top of the hour. Thank you!
Melanie Giuliani to Everyone: Hi everyone - thanks for joining! We will be getting started in just a few minutes.
Melanie Giuliani to Everyone: Alright - let's get started! Please start asking your questions in the chat box now. Please be sure to select "everyone" before sending your questions.
Aditya to Everyone: Can someone briefly explain for which purpose CA UIM is used for? As I am new to this Product I am looking for documentation for CA UIM and I am going to start using this from 22nd of July
Nghia Van to Everyone: is there an effort to allow clustering of the primary nms server? Use case is to run more of an active solution and/or provide ease of scalability
from Dane to Everyone: When is UIM 8.3 going to be released GA?
Klaus to Everyone: NFA Integration architecture?
Jason Allen to Everyone: @Aditya - CA UIM is used for monitoring of IT infrastructure (servers, networks, etc.). Here are a couple of useful links: http://www.ca.com/us/opscenter/ca-unified-infrastructure-management.aspx and https://wiki.ca.com/display/UIMPGA
David LeDeaux to Everyone: @Dane 8.3 is going to be release at the end of the month. We've changed the release schedule a little to help us get more fixes in and give more time for testing.
Jason Allen to Everyone: @Nghia - you can run the primary hub on an active/passive cluster currently - sounds like you're looking more for a "built in" solution with the hub itself for hub clustering, is that right?
Bart Beeman to Everyone: We want to separate prod alerts from lower environment alerts. We are fine with the servers and our naming convention. How would you recommend separating the alerts from the vmware probe?
Martin Kowalewski to Everyone: @Klaus - The NFA Integration Architecture is a NFA 9.3.1 Install (2-Tier or 3-Tier or Singlebox) with NFA Inventory Probe installed anywhere in the environment, it is recommended to be on the NFA Console.
Jason Allen to Everyone: @Bart - it sounds like your best bet will be the NAS AutoOperator rules - you could set up a profile which acts on (assigns, runs a script, etc) all alerts from the vmware probe -- does that help?
Sam to Everyone: Are there any plans to build in a function to enable administrators to extract a report that shows all metrics and thresholds? Working in a managed service environment we are constantly asked by our customers: "What is currently monitored?"
Dane to Everyone: is there a UIM product release roadmap that could be posted to wiki or community? Would be nice to have a determination when new releases were scheduled and even better if this was updated with upcoming functionality
Nghia Van to Everyone: We are looking for a cleaner active solution, run more than one data engine for processing, also really cluser the UMP sites. There seems to be limitations with the clustering setup as CA doesn’t really support clusering of the jasper server etc.
Klaus to Everyone: Thanks. Is there plans for the collectors to report into the uim server instead of a console?
Tom Jacob to Everyone: What is best practice architecture for multi-site HUB? Specifically robot tenancy for HA failover and failback. Specifically seeking input as to if server endpoints must report to nearest local HUB prior to any HA activity for HA functionality to work as designed. In real world environment experienced UMP's not failing back when HA event is over.
Michael Koprowski to Everyone: Are there plans or is there documentation available today for adding reports to Unified Reports? For example availability reports?
Jason Allen to Everyone: @Sam we don't have anything specific on the roadmap at this time, but it is an area we are definitely exploring. In the meantime, the following article might help you generate some of this information: https://na4.salesforce.com/articles/HowToProcedures/How-to-create-a-report-of-all-probe-configurations?popup=true
Michael Koprowski to Everyone: @Sam we I would be interested in similar functionality
Gary Moore to Everyone: I would like to display "Free Disk Remaining" in a dashboard - not as percent, but as actual MB. The cdm probe does not provide that QoS metric. Is there another way to collect that QoS?
Michael Koprowski to Everyone: @Jason the link you provided is asking for a salesforce login.
Kurt Spence to Everyone: @Sam - there are also near term roadmap items coming around policy based management where policies would be able to audit what is being managed. So pls stay tuned for updates on these features coming
LeDeaux to Everyone: @Klaus can we get more explanation on your question regarding collectors reporting into a console?
Jason Allen to Everyone: @Nghia, as far as I know there's nothing specific planned in this area in the short-term, although high availability is something that's always of interest. If you have any specific needs I would definitely recommend raising Ideas in the community.
Martin Kowalewski to Everyone: @Klaus - If you are referring to the Harvesters the data is still stored locallly and the data is presented inside of USM in the interface context.
Tom Jacob to Everyone: @Sam - Related to probe audit. When inquired of similar question was informed that functionality would be coming potentially with 8.3. Also was suggested that 8,3 would potentially include more direct threshold management for end consumer to control thresholds. Can you confirm?
Jim Perkins to Everyone: @Dane: I think what you are asking for is a roadmap with release dates. Unfortunately, we can not publish future release dates.
Marcus Darby to Everyone: I'm currently running the primary HUB on a Win 2012 VM, can I move it to a 2 node active/passive cluster with out reinstalling it?
Jerry Hall to Everyone: We have V-center monitoring and using snmptd to capture the V-center alarms. I have also added the vmware probe to monitor the V-center Virtual Machines. Which of the two above do you recommend (keeping the V-center monitor and capturing traps from it or shut-down the V-center monitor and just use the vmware probe to monitor the Virtual Machines)?
glenn burns to Everyone: @Sam - I've asked for something similar, but no real solution other than try to do it on your own via probe queries and lua scripts
Nghia Van to Everyone: Also using the H probe for failover is not idea as it only addresses the alarm flow, USM does not fail over to the secondary server running the HA data_engine so USM does not fail over and wasp dies.
Klaus to Everyone: Sorry. My question is around NFA integration with UIM. Will the console disappear?
Jim Cooke to Everyone: @Tom Jacob Currently the UMP failover handling is not part of the HA probe scope. People have utilized custom scripts to handle the special needs of UMP failover. We are looking at bringing this capability into the product in the future.
Raj Sundaram to Everyone: @Jerry ..vmware probe would be the preferred choice.
Dane to Everyone: @Jim Perkins - What about a general product roadmap? PBM was release CR 8.2, partner speedcast stated this was not included in 8.3 - just trying to get an idea when we can expect that and other added functionality
Tom Jacob to Everyone: @Jim. Not UMP functionality itself. We are doing that through custom script to change attributes. The robot itself does not fail back even though in packet captures we see it attempt broadcast.
David LeDeaux to Everyone: @Gary I'm looking into this for you currently
Kurt Spence to Everyone: @Michael - you can also try the UIM doc wiki link: https://wiki.ca.com/display/UIM82/CA%20Unified%20Infrastructure%20Management%20-%208.2
Mike Dioguardo to Everyone: I know about the net_connect and icmp probes to determine if a node is unreachable. Is there another actionable alarm that will be generated if the hub or a probe is unreachable without these?
Jason Allen to Everyone: @michael regarding UR reports, we don't have much "how to" documentation but here's a place you might start: https://wiki.ca.com/display/UIM82/Designing+an+Ad+Hoc+View
Sam to Everyone: @kurt Thank you, I'll eagerly await a potential release.
Nghia Van to Everyone: Cloud Monitor integrates with UIM out of the box via a probe. UIM "supports" multi-tenancy but it does not appear that Cloud Monitor does not, any work to correct this?
Michael Koprowski to Everyone: Got it @Kurt, thanks
Martin Kowalewski to Everyone: @Klaus - The Console will not disappear in the short term this is a multiple phase approach. the NFA console provides more than just the web ui. Our long term vision is to remove the console, but it will take time to get there.
Jerry Hall to Everyone: If vmware probe is the best choice, should I shut-down the V-center monitoring? Kinda scared that the vmware probe may not capture everything that V-center does.
Tom Jacob to Everyone: @Jim Cooke - To further clarify. Question is solely directed at robot reaction to HA failover and failback.
Klaus to Everyone: @Martin Thanks
Santhosh Shankar to Everyone: Any update on the capman uim integration ??
Jason Allen to Everyone: @Marcus unfortunately you would need to reinstall in that situation.
Jason Allen to Everyone: @Marcus your best bet might be to set up a new instance pointing to your existing database, then migrate the robots
Kurt Spence to Everyone: @Dane - we are working right now on the GA timing for policy monitoring (not controlled release, but GA). There will be another controlled release end of this month too. So hope to announce shortly
glenn burns to Everyone: UIM to Spectrum integration, are both product team working toward a better integration?
Jerry Hall to Everyone: So, if vmware is the preferred, should I shut-down the V-center monitoring?
glenn burns to Everyone: Any thoughts on Spectrum to UIM integration (2 way)
Ponnuchelvam to Everyone: we are facing issue on cdm probe. it is trigger false alerts for NFS filesystem in Unix environment . We are using latest version 5.40. We are solving this issue by degrade the probe to 4.54. Why latest probe having issue? Do you have any fix?
Raj Sundaram to Everyone: @Jerry...can u send me a list of specific things that you are trying to monitor ... the vmware probe is very comprehensive but we can parse your needs better … please send me a note.
Joe Poutre to Everyone: I too am interested in the UIM-Spectrum integration, as we're going to initiate that connection sometime soon.
Jim Perkins to Everyone: @Glenn, Yes. We are defining a plan presently for the 8.4 release to improve the bi-directional integration.
David LeDeaux to Everyone: @Ponnuchelvam Can you open a support case for this?
Nghia Van to Everyone: When will 1 for 1 functionality be done with Infrastructure Manager and USM as to remove the need to use 2 tools for configuration?
Jerry Hall to Everyone: @Tom - If vmware probe is the best choice, should I shut-down the V-center monitoring? Kinda scared that the vmware probe may not capture everything that V-center does.
Joe Poutre to Everyone: @Jim - what version(s) of Spectrum are supported?
Jason Allen to Everyone: @Nghia, regarding the UMP failover issues, this is something we're working on improving in the product in the future. Regarding cloud monitor integration, I'm not aware of any fix for multi tenancy there unfortunately.
Raj Sundaram to Everyone: @Ponnuchelvam ... have you raised a defect? If not please do so and I will follow up.
glenn burns to Everyone: @Jim - any thoughts on improving integration to provide specific cause codes to each probe?
Venith to Everyone: My requirement would be to create a group and make it visible in USM to multiple user accounts. How can i achieve this?
Kurt Spence to Everyone: @Santhosh - we have UIM/capmap on our roadmap now, and working on that solution. So we plan to announce that in the future, so pls look for those updates
Jim Cooke to Everyone: @Tom Jacob The robot failover and back again is a supported scenario. I do seem to recall an existing defect in some corner case, perhaps dealing with a passive robot scenario. Sounds like we need to engage support to repeat the issue. In most cases the failover works as expected.
Jason Allen to Everyone: @Nghia we are targeting the UIM 8.4 release for full IM replacement.
Ponnuchelvam to Everyone: We have raised case against this issue. We are waiting for fix
Jim Perkins to Everyone: @Glenn, Yes. But I don't know if we will have that in place for the 8.4 release. We are calling it the "Event Catalog"
Nghia Van to Everyone: Plans to standardize basic functionality across probes? for example customize message pools, subsystem IDs etc.
Jason Allen to Everyone: @Nghia just to clarify - I assume you mean replacing IM with Admin Console
Nghia Van to Everyone: @Jason yes
Jim Perkins to Everyone: @Joe: This will likely be only 10.1, but it is too early to say definitively.
Jason Allen to Everyone: @Nghia ok thanks. That should be coming in 8.4. We do have some plans to bring some more administration capabilities into the USM portlet but that's longer term.
Marcus Darby to Everyone: Thanks for the responses. I'm new to UIM but it don't seem like using the HA probe and promoting a second HUM for failover is a good solution for HA.
Santhosh Shankar to Everyone: Any documentation on the IT Health Index feature ... We are missing use cases to sell this better
Dane to Everyone: @Nghia Van hopefully Admin Console is enhanced to support all probes, all IM functionality - and isn't so dog slow
glenn burns to Everyone: @ Jim, thanks for the update. Any change we can get involved in that process. We are currently going through a migration from legacy tools to UIM->Spectrum and we have lots of concerns with the current integration
Kurt Spence to Everyone: @Nghia - in the past couple releases there has been work on a standardized block within the admin console for probe config. This is still work in progress and we expect to continue refining this in the 8.3 and 8.4 timelines
Melissa Potvin to Everyone: @santhosh: also we have a community dedicated to capman. You may want to follow if not already...
Sam to Everyone: Does the "disk capacity prediction" functionality work on 7.6?
Deb E to Everyone: Any thought to having pre-processing rules be hierarchical or at least go through all that match?
Timothy O'Connor to Everyone: @Venith We'd like to change groups to be able to belong to multiple accounts but this isn't planned for the near future. We are working on having users belong to muliple accounts which would allow a user to see all of the groups for those acounts.
Nghia Van to Everyone: @Jason, any documentation on those plans? I have to maintain a road map for our solution and inheritly im coupled to CA's roadmap
Kurt Spence to Everyone: @Sam, for disk cap prediction are you referring to the time to threshold? This was released in 8.0
Jerry Hall to Everyone: We have a customer in India who is concerned with Latency and Bandwidth between their Secondary tunnel Hub and our Primary tunnel Hub in the US. Can you provide a ball-park idea on the bandwidth required between the two Hubs if they have 500 servers being monitored and if Latency is even a concern?
Sam to Everyone: @kurt yes, that's the one.
Kurt Spence to Everyone: @Sam, you'll need to be at 8.x for time to threshold
julio to Everyone: @Jim.. so, when are you planning to have the "Event Catalog" available? We are in the middle of an integration between UIM and Spectrum and we need this functionality to be able to correlate alarms better
David LeDeaux to Everyone: @Gary, I'm not finding anything quickly regarding disk free % vs MB. Any chance you can open a support case for this and we can investigate a little more in depth?
Jason Allen to Everyone: @Nghia I'm not aware of anything we have at this time but I'll discuss with product management and see what we can get for you.
Sam to Everyone: @kurt, thank you.
Jim Perkins to Everyone: @Deb E. Look for the new Event Management Service in 8.4. It will have an improved rules and matching engine when released.
Jim Cooke to Everyone: @Marcus Darby. We have a longer term roadmap plan to replace the current long standing HA probe failover practice, with standard OS level 3rd party failover solutions. But this is still quite a way out for us.
Sam to Everyone: Are there any plans to enable more variables to the Triggers? The current selection is limited.
Venith to Everyone: @Timothy: Earlier this used to work in Dynamic dashboard view, so we are looking for a similar console.
Kurt Spence to Everyone: @Sam, also note that for CDM specifically as of now the predictive alarms are only on dynamic (baseline deviation) alarms - time to thresh for statics is still coming on the release roadmap
Nghia Van to Everyone: Any plans to support jasperserver clustering within USM? USM can run in a clustered manner behind a load balancer but jasper server can only run on 1 server and be supported by CA currently
Jim Perkins to Everyone: @Julio It will be part of EMS in 8.4, but I don't know if it will give you what you need. Are you looking for a combined catalog between Spectrum and UIM?
Kurt Spence to Everyone: @Sam are you referring to triggers for NAS auto oper?
Gary to Everyone: Is it possible to map an LDAP group to a site within the UMP instead of having to doing this manually for each user?
Sam to Everyone: @kurt, yes. That's them.
Kurt Spence to Everyone: @Sam no plans as of today to enhance NAS auto oper - can you briefly detail your use case? maybe another way...
Jason Allen to Everyone: @Jerry, unfortunately we don't have any specific documentation on bandwidth requirements - but I can tell you that the average QoS message is about 2kb. So, if you can determine the number of total QoS metrics you might be monitoring, and the average monitoring interval then you could ballpark this. Example: 500 servers, sending 10 QoS metrics each (5000 metrics) every 5 minutes - so you'd need 10000kb of bandwidth every 5 minutes, or 2000kb/minute, which equates to about 33kb/sec for that level of monitoring. Given that the data would likely be spread out over the entire interval you could potentially get away with less than this
glenn burns to Everyone: @Jim - julio and i work together. we are looking for a better correlation and integration. having the same cause code in Spectrum regardless of the alarm type or probe type is very prohibitive when it comes to alarm enhancements and spectrum correlation
Sam to Everyone: @kurt, sure. Ideally the same variables available in the NAS e.g: $hostname, $message
Jim Perkins to Everyone: @Sam. Yes. We are releasing the Event Management Service with 8.4 with a more flexible method for matching on any event/alarm property.
Timothy O'Connor to Everyone: @Nghia Regarding jasperserver clustering. I'll need to look into that and get back to you. Can you send me your email privately and I'll follow up?
Marcus Darby to Everyone: So you agree that windows clustering would be the better solution in my case for now? My database is already clustered and at this time I'm not addressing site failover.
Sam to Everyone: @kurt, sure. Similar variables as the NAS, $hostname $message etc
Santhosh Shankar to Everyone: Do we have any plans to broaden our coverage for Nosql databases apart from Cassandra , Mongo and Hadoop ?? What’s the strategy and message to customers who use more than 6 open source NoSQL DBs ??
Jason Allen to Everyone: @Gary unfortunately each user is required to log into UMP one time to have a user account created to associate with a site, so it's not possible at this time.
Joe Poutre to Everyone: @Jim - we have the same situation with our SysEdge alarms into Spectrum. I've had to work at creating custom events in Spectrum so as to properly portray and route such alarms. I would prefer not to have to do that again for the UIM integration.
julio to Everyone: @Jim.. No… we are not looking for a combined catalog just a better integration and more flexibility with cause codes so that we can take full advantage of Event Configuration in Spectrum
glenn burns to Everyone: @jim, ideally having different cause codes for different types of probe makes it simpler to be able to correlate
Nghia Van to Everyone: @Timothy I have no way to sending you my email address privately but you can ping me at ******@xxxx.com
Venith to Everyone: We have observed inappropriate interface utilization value using interface_traffic probe for few Firewalls and Switches. Version downgrade have fixed them. Does the latest version has a bug ?
Timothy O'Connor to Everyone: @Nghia Thanks. I'll see what I can dig up and get back to you.
Paul Breheny to Everyone: @Venith, I don't know of a bug like that, I recommend opening a support case on that issue to look at the particulars.
Jim Perkins to Everyone: @Julio and Glenn. I would like to talk to you personally to go over the details. Can we speak directly off-line? I will send you an email to reach out.
glenn burns to Everyone: @ jim - Sure
julio to Everyone: Any ideas of allowing to create "Groups" within the IM or admin console and be able to deploy packages ?. We have a need to deploy packages to a group of servers running a specific app.. can't filter using robot name
Raj Sundaram to Everyone: @Santosh .. yes we are planning to add more big data/noql probes .. some wi ll be CA built, others will be partner contributed via the probe marketplace.
Jerry Hall to Everyone: @Glenn - Thanks, any idea if Latency between India and US should be a concern? Are other customers doing this with no issues?
Kyla to Everyone: Any plans to allow the auto-operator to be able to use OS information from device alarms in profiles, triggers, etc.? Ex. Setup an email operator profile to filter alarms from Windows devices.
julio to Everyone: @Jim.. Sure... just send us an e-mail
Nghia Van to Everyone: Plans to truly make USM multi-tenant? I can create customer communities but if one custer creates a group called "Windows" all of a sudden no other community can create a group by the same name
Gary Moore to Everyone: @David - Thanks, I have already opened a case (suggested I open an idea), and know that there is already an idea posted to add the Free Disk Space Qos to the cdm probe. I was hoping someone had come up with a workaround I could use.
Venith to Everyone: Regarding ZONE monitoring : I have a customer running his platform in Solaris LDOM with zones and we observe the memory values in server and the one cdm probe fetches do not match. The critical alerts contains the Memory percentages as (2081%) and could also see the value dropping to negative in clear alerts
Jim Perkins to Everyone: @Joe. I'll reach out to you too. I want to make sure I am catching all of these integration use cases for our requirements.
Jim Perkins to Everyone: @Joe. I'll send you an email
Joe Poutre to Everyone: @Jim - Thank you, I'll expect it.
David LeDeaux to Everyone: @Gary can you send me the case #? I'll follow up
Raj Sundaram to Everyone: @Venith .. have you raised a case for this ? If not please do so and advise, we will follow up.
Jim Perkins to Everyone: @Kyla: Yes. The new Event Management Service will have capabilities that support that. Look in 8.4
David LeDeaux to Everyone: @Julio We're implementing a feature called Policy Based Management for deployment, but it is still largely based on what we know about a device -- OS, Name, IP, etc. We don't have a good way to know which applications are installed on a server. I believe we expose the User Tags to this, however so that might be an option.
Timothy O'Connor to Everyone: @Nghia Unfortunately group names will collide across tenants. We'd like to address this but it isn't in our short term plans. Have you submitted an idea for this?
Venith to Everyone: @Paul: The high interface usage alerts were mostly for Tunnel interfaces in Firewall.
Nghia Van to Everyone: I know CA is putting great effort into trying to fix/enhance alarming but is there an equal effort to look at QOS? for example QOS cleaner QOS data merging/cleanup via USM? Maybe the ability to support data warehousing?
glenn burns to Everyone: @David - Policy Based Management was supposed to be released with 8.2, we know its not in 8.3. any ideas when the new release date for Policy Based Management?
Klaus to Everyone: Is there any future plans for configuration management for routers and switches or is that going to stay in Spectrum?
julio to Everyone: @David.. having the user tags as part of it should help... I just need to be able to updated user tags to hundreds of servers at once
Eyra to Everyone: has any have an issue with database performance? I mean the nas probe saves null
Kurt Spence to Everyone: @Marcus, we have some standard fail over reference arch internally that our consultants use with our customers. I'll reach out to you after and go over that
Venith to Everyone: @Raj Sundaram: We already have a case open for this xxxxxxxx and its open for the last one month.
Kurt Spence to Everyone: @Eyra - are you seeing DB issues?
David LeDeaux to Everyone: @Glenn We'll need to have someone from Product Management answer that
Eyra to Everyone: @kurt well It seems that it has a lot of fragmentation I mean the indexes of the tables but the problem is that I don’t know why the indexes are on that way
Marcus Darby to Everyone: Thanks Kurt, I appreciate and look forward to it.
Eyra to Everyone: @kurt and when I review the index that has 94% of fragmentation it corresponds to a table of nas probe
Venith to Everyone: Thanks. It would be helpful if we have a fix for the same at the earliest.
Jim Cooke to Everyone: @julio: We have stories in our backlog to support creating groups in Admin Console which could be used to easily deploy probe and updates. We are not sure yet if this will make the 8.4 cut or not. We are also improving the search ability to identify a set of robot you are interested in.
Jerry Hall to Everyone: When setting-up the HA Failover server, should I add all of the same probes to it as I have on the server I'm failing-over from?
Kurt Spence to Everyone: @Glenn -PBM is in controlled release now (as of 8.2) and will be updated again as such end of this month. I can reach out to you after this for how you can start with it
Eyra to Everyone: @kurt the curious think is that the name of that index is "NULL" and if you looking on the table names "NAS_SUMMARY" it has a lot of null values
Gary Moore to Everyone: @David - Free Disk Space Case Number xxxxxxxx
Kurt Spence to Everyone: @Eyra - can you open a ticket? It'd be good for support to check out what you're seeing? can't say as of right now
Eyra to Everyone: @kurt ok =)
Nghia Van to Everyone: I know CA is putting great effort into trying to fix/enhance alarming but is there an equal effort to look at QOS? for example QOS cleaner QOS data merging/cleanup via USM? Maybe the ability to support data warehousing?
Venith to Everyone: Regarding new probes development : I am looking for the right procedure to start of with new probe development, do we have a details document for the same, Also i would like to know, if there are any options to breakdown and modify an existing probe.
Melanie Giuliani to Everyone: Hi @all - this is the 15 minute warning! Please get your last minute/follow-up questions in now!
glenn burns to Everyone: @Kurt, that would be great
Jim Cooke to Everyone: @Jerry Hall: It is not requirement to match the monitoring probes on the HA failover system. Many people do not do this, and only configure the critical probes. This was also the original thinking behind why UMP was not included in the HA failover, because during this time it was not originally perceived as critical.
David LeDeaux to Everyone: @Eyra If you're seeing high fragmentation levels, we recommend having a SQL maintanence plan in place to defragment indexes. Here is an article related to that: https://na4.salesforce.com/kAH6000000000NF?srPos=2&srKp=kaH
Pete to Everyone: @Klaus. Yes we are currently working on a project to provide NCM like functionality to manage change and configuration of switches, routers etc. If you would like to discuss in more detail under NDA please contact me at ******@***.com and I'd be glad to discuss in more detail.
Klaus to Everyone: @Pete Thanks
julio to Everyone: @Jim I wasn't aware you could create groups in Admin Console.. I will look into it. I was only able to create groups in IM but my understanding is that "Groups" is a deprecated concept in IM
David LeDeaux to Everyone: @Eyra Bear in mind that some tables will appear to be fragmented constantly because they have very little data in them. It's more efficient for the SQL engine to load a small table into memory to scan it than to reference an index.
Kurt Spence to Everyone: @Nghia - we do have data engine enhancements on our roadmap. And we also working on APIs for direct access to data for customers wanting to warehouse - we can reach out to you on plans for that coming soon. What else would you want re: your comment of merging of data? (vs what we do now for raw/hour/day roll ups?)
Klaus to Everyone: @Pete Will it include policy violations?
Jerry Hall to Everyone: @David; Other than the monitoring probes (i.e. cdm, process, etc.) what probes are required on the Fail-over Hub?
Jim Cooke to Everyone: @julio: This is a FUTURE Admin Console capability, not in the current GA release.
Eyra to Everyone: @David ok, that helps me to keep low fragmentation of the indexes, but... why (in general) we could experiment high level of fragmentation?
Gary to Everyone: Can configuration and super packages be deployed from the Admin Console?
Pete to Everyone: @Klaus, yes we have Features and user stories on policy enforcement.
Nghia Van to Everyone: @Kurt Clean up of QOS for example use case 1 if a server gets renamed because of naming convention changes with customer we should be able to easily migrate the old qos data to the new name
Klaus to Everyone: @Pete Thanks. I will contact you via email
Jim Cooke to Everyone: @julio: This Admin Console ability to establish groups for managing monitoring infrastructure is already on the on CA Communities Idea Forum and in our backlog.
Sam to Everyone: @nghia I like that idea
David LeDeaux to Everyone: @Jerry As a 50,000 ft view, my recommendation when setting up an HA hub is to completely replicate the behavior of the hub that you are creating a secondary for. I like to actually bring up a secondary hub and take down the primary and configure it. There's a lot of nitty gritty that comes into the specifics
Jerry Hall to Everyone: @Glenn - Thanks, any idea if Latency between India ans US should be a concern? Are other customers doing this with no issues?
Kurt Spence to Everyone: @Nghia, yes that is a common case. Today we have call backs to remove inventory from the system for no more collection. Then we have roadmap feature planned for also allowing data grooming of removed devices, so that's coming
julio to Everyone: @jim thanks
David LeDeaux to Everyone: @Eyra Fragmentation in an index is expected. It's a lot like making an appendix in the back of a book and just adding new entries to it. They won't be in order until you rewrite the appendix to be alphabetical. This is made worse by inserting and deleting data.
Venith to Everyone: Regarding NAS: Do we have any limitation on the no. of rules (AO and Preprocessing) that can be created in nas. Also is there an option to move a rule up and down by setting an order value??
Jim Cooke to Everyone: @Gary: Existing superpackages can be deployed via Admin Console (using ADE under the covers). The ability to create new super packages is something we are working on for the new Admin Console planned for 8.4.
Nghia Van to Everyone: Also certain probes only report via IP address (cisco monitor off the top of my head) soe other probes do other things and then we end up wit multiple objects in USM for what is actually 1 device, there should be a way to group these objets and tell UIM that this is actually 1 devices and it sticks
Nghia Van to Everyone: @Kurt can you send me your road map even if it is just ball parked?
David LeDeaux to Everyone: @Venith I'm not aware of a specific limit; do you have a ballpark figure on the number that you are implementing? Also, you are able to sort the auto operators currently
Kurt Spence to Everyone: @Nghia I'll follow up with you afterward on how we can do a futures discussion
David LeDeaux to Everyone: @Venith The sort order can be changed by Right clicking the rule and selecing Move Up / Move Down
Klaus to Everyone: I seems not finding any courses for release 8 or later Only for 7.6?
Nghia Van to Everyone: @kurt thank you... I have to warn you I have lots to say
David LeDeaux to Everyone: @Venith there's also a complimentary rule within auto operators on the advanced tab to stop processing further rules on a match
Melanie Giuliani to Everyone: @all - we have 5 minutes left! Please ask any last-minute follow-up questions now. We will be closing out the meeting at 12pm ET.
Melissa Potvin to Everyone: @Klaus, I will follow up with our CA Education team and will get back to you
julio to Everyone: @David.. is that order in NAS also consider to be a Priority?. If I want certain rule to be hit first
Jim Cooke to Everyone: @Nghia Van: We are constantly working on improving the Discovery Server algorithms for correlating devices reported from different sources so they can reconcile to the same device. We have future plans to externalize these rules so the precedence ordering can be manually manipulated.
Nghia Van to Everyone: Plans to allow excel like imports to set up maint mode? sometimes we get a list from the customer with a handful of devices and soetimes it can literally be hundreds of devices with outage times all over which is tedious for the NOC with the current maint mode
David LeDeaux to Everyone: @Julio yes, it will execute the top rules first
Klaus to Everyone: @Melissa Thanks
Eyra to Everyone: @David LeDeaux.- ok thanks for the explanation =) so.... I have a lot of doubts and problems ha! the first and I mean the most important is the database perfomance, the sencond one is that the dashboards that I design, well they simply don't publish any data....
julio to Everyone: @David.. how about a pre-processing rule? is there a way to prioritize them?
Venith to Everyone: @David LeDeaux: Currently we have 575 AO rules and 255 Pre-processing rules in our environment... and i would be able to sort Ao rules, but changing the order could make the rules to misbejave since we have several customers critical and clear profile configured
Jerry Hall to Everyone: Is the snmpcollector probe good to use now (I'm v8.2) or should I wait until v8.3?
Tom Jacob to Everyone: @Jim Cooke. May I email you regarding the HA topic from earlier?
Jim Cooke to Everyone: @Nghia Van: This maintenance mode import idea is interesting. You should post it on the CA Community Forum for voting. There is a maintenance mode API – maybe that can you be used for you case.
David LeDeaux to Everyone: @Eyra We have some hotfixes available for 8.2 that resolve some dashboard issues. Are you able to open a support case?
Sam to Everyone: @nghia I'd vote for that!
Kurt Spence to Everyone: @Eyra, and do you have a case open for any DB perf issues?
Jim Cooke to Everyone: @Tom Jacob: Yes you can email me privately.
Nghia Van to Everyone: well we created our own maint mode and it works just fine and supports imports....i just was hoping CA would make it work
Jerrod Early to Everyone: Is there any possibility of a complete re-write of the sqlserver probe in the future? There are huge gaps in what is monitored. I have seen some other DBAs frustrated with the probe as well, so I just am wondering.
Eyra to Everyone: @kurt this is the case of null data on table of NAS probe: xxxxxxxx
Melanie Giuliani to Everyone: Alright, that's it for office hours today! No more new questions, please, but we will finish answering any existing questions/follow-ups. Thank you everyone for joining! We will be posting a transcript of this session to the IM community later today. If you aren't already a member, please be sure to follow the IM community for updates and more UIM events!
Kurt Spence to Everyone: @Eyra we can follow up on your case
Daniel Okine to Everyone: @Jerry Hall. You should be able to use snmcollector in 8.2 re there any particlaur issues preventing you from using it> Of course we are incrementable enhancing it., with the new release coming out with 8.3
julio to Everyone: @Nghia.. are you able to put a specific probe in maint? not the robot.. right?
Eyra to Everyone: @kurt see you there!
Sam to Everyone: Great meeting, very interesting. Thank you.
Nghia Van to Everyone: @yes i have granularity where i can act on probes vs robots in my custom maint mode
Venith to Everyone: Hope i would be getting a response for this question : "Regarding new probes development : I am looking for the right procedure to start of with new probe development, do we have a detailed document for the same, Also i would like to know, if there are any options to breakdown and modify an existing probe."
Jim Cooke to Everyone: @Julio @Nghia: I was speaking to the USM based Maintenance Mode capability introduced with 7.5 (I think), not the old legacy robot based maintenance capability.
julio to Everyone: @Nghia.. I would like to hear more about it.. we are currently looking for something like that where we have a need to put a specific probe or even a profile within the probe in maint
Nghia Van to Everyone: @julio you can email me at xxxx@xxxx.com
Eyra to Everyone: we could open a chat using IRC
julio to Everyone: @Nghia.. Thanks I will
Melissa Potvin to Everyone: That's a wrap for today. Thank you for joining!
Kurt Spence to Everyone: @Venith - I'll try and get some info to you on probe dev
Nghia Van to Everyone: thank you everyone toodles
Gary Moore to Everyone: Thanks all!