DX Infrastructure Manager

Expand all | Collapse all

**** Defect Announcements ****

Anon Anon03-19-2015 08:03 PM

  • 1.  **** Defect Announcements ****

    Posted 11-13-2014 11:11 PM

    **** Defect Announcements ****

    Since the Announcements section is not used on the official support page I'll create one in here where most folks check.

     

    I'll keep this thread going with new finding, bugs, defects that I find in the GA probes so others will become aware of issues:

     

    NOTE:

    If posting a new defect please include the following information regarding the defect so others can quickly identify if they would be effected by the defect you found. Please start the post off with:

     

    UIM Version #.#       Exp:   UIM 7.5

    UIM Component:        Exp:  UMP, HUB, Portlet - USM, SLM, etc...

    Probe_Name v#.#       Exp:  CDM 5.10

    OS Running:           Win|Lin|Unix

    Description of Defect:

    TAGS:  <key words regarding the defect, include the probe name if probe is part of the defect. Very helpful as this will allow others to find the defect when searching.>

    Ex: cdm, breaks, threshold, false, alert, hub, slow, etc....

     

     

    So here to start off:  

    ------------------------------------

    UIM:  7.5 / Robot Base v7.62 Build # < 1330  Fixed in Latest Robot Build 7.63

    Bad robot build of v7.62.  Found 10/22/14

    If you have Robot version 7.62 and the build is not 1330 you really need to upgrade to the latest build 1330 and re-deploy. The issue was whenever the robot went down, machine rebooted, service stopped the robot would automatically un-register itself from the hub and thus you would never get an Robot is inactive alert. It would just remove itself from under the HUB.

    FIX Solution: 12/12/14:  Update Robot to latest version with Build >= 1330.

    ------------------------------------

    UIM 7.5, Probe: NTEVL v4.01  HIGH CPU Usage 10/30/14 - Fixed in ntevl v4.02

    With our current ntevl v4.01 if you have many exclusions in your ntevl profile, with this latest version it eats up CPU cycles on the machine. Downgrade to v3.84. I had another issue with v3.90 where I had bad memory leaks on Exchange Servers running the 3.90 ntevl probe. So far v3.84 seems to be w/o issues for me.

    2/5/2015: Latest version of NTEVL 4.02 CPU usage finally is back to normal.

    ------------------------------------



  • 2.  Re: **** Defect Announcements ****

    Posted 11-14-2014 09:04 AM

    There are a lot, but one from this morning:

     

    3. ntservices (at least 3.12 and 3.15) can't handle services that have < > in the service name, and text after that. They appear in the profile list, but you can't activate them (even through raw config)

     

    -edit: actually fixed in 3.16, though updating doesn't solve it without deleting some of the service sections

     

    -jon



  • 3.  Re: **** Defect Announcements ****

    Posted 11-14-2014 06:06 PM

    Also:

    Bad build of the robot_exe v 7.62 - using with bulk deployment, third party (SCCM), the robot will not register as service

    Bad build of the robot_aix v 7.62 - using with ADE, will receive net.lingala.zip4j exception saying file headers not found



  • 4.  Re: **** Defect Announcements ****

    Posted 11-19-2014 10:05 PM

    Hub 7.61 

     

    Client request over a tunnel receives a TCP/RST and the connection is closed before data transfer of response finishes under
    specific circumstances. NIME_COMM and nimSessionGetLastError=104 filter through to the requester and the TCP connection is reset by the hub before completing transmission. Gui config tools usually report this as a communication error. The hub on the request side logs cadence check SSL, while the remote end logs successful transmission of response at high debug levels.


    The trigger seems is extra latency in the network on the response receiving end of the tunnel along with a large response to a callback.  IE: requester<VPN<hub<tunnel<hub<destination.  Tunnels got faster in 7.61 which flushed out the related bug that only appears when latency on the last hop of the response is greater than latency accross the tunnel.

     

    Fixed in a 7.62 hub beta is available through support if you run into it. Last report was that it missed a cutoff for the next GA, but will likely make it into a GA release in January.



  • 5.  Re: **** Defect Announcements ****

    Posted 11-20-2014 01:35 PM

    Hi

     

    CDM probe version 5.02 on AIX causes multiple defunct process.Support requested us to downgrade to 4.91.

     

    Regards

    Mehmet



  • 6.  Re: **** Defect Announcements ****

    Posted 11-20-2014 11:37 PM

    I found this one out today. Don't upgrade to the latest version of the IM Tool.

    I thought someone removed the right to the Account Administraion GUI. Under the Security Menu in the IM Tool, there is no longer the "Account Administration" entry. 

    I opened and issue and support engineer was also "surprised" that product development removed the feature from the IM Tool.

     

    Of course no release notes stating this. So stay on the 4.08 version. I upgrade to v4.10 and thought someone removed my right but no. It was UIM who removed this feature. Their response was we now have to perform this function within the UMP Account Administration portlet. 

     



  • 7.  Re: **** Defect Announcements ****

    Posted 11-21-2014 11:20 AM

    Hub 7.61 no longers generates an alarm if a tunnel client cannot connect to a server.

     

    Here's their explanation:

    "Due to major architectural changes in the way tunnels work in hub 7.61 compared to previous versions, the tunnel disconnection alarms were no longer reliable - false alarms would be generated very frequently as a result of these architectural changes. Therefore, product management made the decision to have the development team remove these alarms entirely from the hub -- alarms will no longer be generated upon simple tunnel disconnection."

     

    Of course, none of this is mentioned Release Notes/Revision History.

     



  • 8.  Re: **** Defect Announcements ****

    Posted 12-09-2014 10:59 PM

    alarm enrichment 4.40 breaks alarm message encoding.

     

    If you send an alarm with special characters, for example umlauts (ö, ä etc), alarm enrichment breaks them.

     

    Defect DE38847 has been raised for this issue.

     

    -jon



  • 9.  Re: **** Defect Announcements ****

    Posted 12-10-2014 12:22 AM

    Logmon - 3.45 failed to start thread too many retries
    Description In trying to troubleshoot a different issue with logmon we upgraded the probe from 3.25 to 3.45. The gui / communication issue still existed (you will see this in the recording) but we also noticed that the probe was not monitoring. After upgrading the probe we started to see this message in the probe:


    Nov 5 12:30:54:726 [140156860262176] logmon: failed to start thread
    Nov 5 12:30:55:334 [140156860262176] logmon: Too many retries when waiting for

     

     

    Looking at the setup section differences between 3.25 and 3.45:

    3.45:

    <setup>
    logfile = logmon.log
    logsize = 5000
    cfg_ver = 2
    debug = 1
    format_interval = 1
    severity = 5
    commOnRunThreads = 200
    winDefaultEncoding = UTF-8
    </setup>

     

     

    3.25:

    <setup>
    logfile = logmon.log
    logsize = 5000
    cfg_ver = 2
    debug = 1
    </setup>

     

    In 3.25 - there is no key called commonrunthreads. In looking at the CA published articles I see this for the URL_response probe:

    "The limit can be modified with the 'max_threads' setting in the configuration file. Be aware that you will need to check the resource usage of the probe when doing so to make sure not to run into machine-dependant limitations for memory, threads or connections. The probe initially allocates 'min_threads' for profile execution. If there are more profiles defined than this, the probe may gradually increase the number of threads. As I mentioned, profiles that not immediately executed are allocated a thread are rescheduled 20 seconds later. You can configure the min_threads and max_threads options via Raw Configuration for the probe by selecting the probe and holding down the SHIFT key, Rt-Click and select Raw Configure. "

    You will notice in the recording that I am not getting alerting when the thread count has been exceed, according to the article above the URL_response probe behaves in this manner.

    The three defects I see here are as follows:

    1) Yet AGAIN - the probe had some undocumented change that broke monitoring and added a new key that is not called out in the documentation. The developers just hard coded some random value (100) and called it good

    2) There is no alert for when this condition occurs. So basically the monitoring that used to work in 3.25 had been broken by the new "feature" and there is no alert. You can see this in the recording

    3) There is no way to set or override this value from IM - which mean you have to log onto the box and manually change this setting within the logmon configuration file by hand.

    NONE of these three items are ideas but rather defects. I will upload the recording and configuration file we are using.



  • 10.  Re: **** Defect Announcements **** raw configure corruption

    Posted 12-11-2014 06:02 PM

    Within config file format, "/" characters in sections/elements are replaced by ## in several places for transport an storage.  On the backend, a flattened version is managed by robots where sections are / delimited as well, so swapping out the potentially ambigous chars is needed.

     

    Raw configure does not do this properly when a slash exists on the key side of an attribute.

     

    A config with:

     

    <mysection>

        /some/key = value

    </mysection>

     

    Becomes

     

    <mysection>

        <some>

              key = value

        </some>

    </mysection>

     

    This has existed for some time and will likely not be addressed given the rarity of slashes in keys, the limited use of raw config on supported probes, and the intent to remove inf manager and related programs like raw configure.



  • 11.  Re: **** Defect Announcements ****

    Posted 12-11-2014 07:41 PM

    I ran into another bug the other day with UIM v8.0 and running the (Windows Infrastructure (HUB, distsrv and Robot) installation from Build 803) DMZ Setup Wizard on a 2012 R2 machine.

    The hub portion installed fine, but when it got to the DMZ Setup Wizard the 8.0 version would constantly crash. I ran it 2-3x and crashed every time. 

    When I ran the 7.6 version it worked perfectly fine. Didn't open an issue yet but just wanted to throw this out there as a FYI.

     

     



  • 12.  Re: **** Defect Announcements ****

    Posted 12-11-2014 08:18 PM

    Hi Dblanco,

     

    I've seen this issue as well. I'm not quire sure, but I have feeling it hasn't occurred on every 2012R2, though. Never investigated it further, usually when I come across such problems I just install IM.

     

    -jon



  • 13.  Re: **** Defect Announcements : (

    Posted 12-13-2014 02:35 AM

    Shopping cart icon in admin console takes you to a fancy looking splash page that insinuates a neat something will be their with fancy marketing letters over a shopping cart saying saying "FALL OF 2014!"  When you click on the link it takes you to the ca home page.  12/12/14 and counting on v8 installed in November.

     

    Now that fancy logo and font is an advertisment for a missed delivery date. It will be even funnier in January.  : (

     

    http://www.ca.com/us/shop/probes.aspx



  • 14.  Re: **** Defect Announcements : (

    Posted 12-13-2014 09:11 AM
    Uim 8 has troubles with queues when hostnames are long. There's a hotfix https://na4.salesforce.com/articles/CurrentHotfixes/Hub-queues-don-t-get-created-or-work-properly-after-upgrade-to-UIM-8-0?popup=true which isnt mentioned in known issues, or fixed in uim setup..

    -jon


  • 15.  Re: **** Defect Announcements : (

    Posted 12-15-2014 05:48 AM
    Thanks. How long of hostname are we talking here? How many characters long can the hubs handle?


  • 16.  Re: **** Defect Announcements : (

    Posted 12-15-2014 09:28 AM

    I believe it only pertains to the hostname of the hub robot itself. The article unfortunately didn't mention, but after updating those two probes, the problem is fixed.

     

    -jon



  • 17.  Re: **** Defect Announcements : (

    Posted 12-16-2014 10:56 PM
    Bugs :
    On Agent startup the agent will always give you the following error this is due to it trying to send a sid and login before doing a nim login :

    Controller: (secCallVerifyLogin) request verify_login failed 66.216.111.168/48002 (permission denied)
    Controller: verify login - cmd=probe_list frm=127.0.0.1/52271 failed

    CDM Bug / Feature :

    cdm currently supports only up to 64 cpu's and the wmi calls has since changed on windows 2k8 and above for more then 64 CPU's

    Cluster probe :
    The probe only supports up to 32 char hostname anything above the probe starts to fail

    HUB & Agent :

    When the hub is set to compatibility mode or SSL only mode you can no longer log into the HUB . Earlier version before 8.0 The client agent if it had snmp probe the snmptd probe and controller may hit 100% cpu utilization

    PCI controller issue :

    The default settings on the agent if not utilizing ssl mode or compatibility mode is not PCI compliant and the agent still generates a weak cert which does not pass a PCI compliance test.

    Controller & Hub Poodle vulnerability :

    Currently as of release 8.0 openssl 1.0c is utilized which has several known security holes.


  • 18.  Re: **** Defect Announcements : (

    Posted 12-23-2014 08:43 PM

    snmptoolkit 



  • 19.  Re: **** Defect Announcements **** ironic ppm infinite recursion suicide.

    Posted 01-08-2015 11:14 PM

    Ppm's job is primarily to get ctd data for adminconsole as far as I can tell.  It's hardly documented.  

     

    As of ppm 3.03 UIM 8.1 and possibly earlier, ppm will repeatedly delegate getting the ctd information to itself until it exhausts it's allocated memory or exhuasts the sockets available to it.

     

    The trigger is selecting configure from the admin console for the ppm probe. Adminconsole requests the ctd info from ppm which requests it from itself ...  You'll see the results in the log and netstat.  Eventually admin console will timout the request, but ppm will still keep going like the little engine that sucks.



  • 20.  Re: **** Defect Announcements ****

    Posted 01-13-2015 09:08 PM

    UMP 8.1 installer deploys dashboard_engine and dap legacy probes to the UMP server by default with dashboard_engine enabled and dap disabled.  For some reason the dashboard_engine subscribes to all qos from the root hub, possibly a change from older versions or a behaviour that happens in the absense of a running dap or variable server.  In a large environment, the dashboard_engine falls behind on the subscription and backlogs huge amounts of qos on the root hub which will fill the partitiion if unchecked.  Dashboard_engine and dap are only to support deprecated legacy dashboards.

     

    Changing java memory settings and enabling dap and/or variable server might fix the issue if you need legacy dashboards.  Disable it if you don't need legacy dashboards.



  • 21.  Re: **** Defect Announcements ****

    Posted 01-14-2015 03:36 PM

    You can also set the following key in your dashboard_engine config:

     

    <data>
        qos_subscription_enabled = 0
    </data>

     



  • 22.  Re: **** Defect Announcements ****

    Posted 01-14-2015 07:59 PM

    when you plan to upgrade to UIM8.0 or UIM8.1 and you are using an oracle backend make sure you also run (despite what is already documented in the release notes)

    grant create type to <db_owner>;

     otherwise your upgrade will fail.



  • 23.  Re: **** Defect Announcements ****

    Posted 01-22-2015 02:01 PM

    controller 7.63 and at least net_connect 3.04 and net_connect 3.05:

     

    removing the probe doesn't work properly. When you delete it, ithe status turns yellow and the port unregister seems to fail.

     

    -jon



  • 24.  Re: **** Defect Announcements ****

    Posted 05-19-2015 04:02 AM

    I tried this with controller 5.70 and it happens there as well.

    So far backline has informed that this doesn't happen in Admin Console and is a flaw of IM and wont be fixed as it'll be deprecated. However, since the port is not removed from the controller, and it doesn't show in AC in that state at all, it seems to me that this is actually an additional defect of the AC.

     

    As a workaround, after removing net_connect, you can manually call port_unregister with parameter name = net_connect to remove it.

     

    Case still open..

     

    -jon



  • 25.  Re: **** Defect Announcements ****

    Posted 01-22-2015 05:51 PM

    ntperf/ntperf64: Comparisons don't do anything nor trigger any alarms in the ntperf probe. 

     

    ntperf / ntper64 probe v1.86-*1.89 *edited: tested with the latest version that was released today 1/22/15 and still broken Did anyone test this probe?*

    Created a simple WMI query on Memory Usage on the box the probe was deployed. Grabbed the current Memory Usage value and set alert to Generate an alerts if != 500. Apply and no alarm getrs generated. Log shows the values do not match yet no alarms is triggered. basic probe functionality doesn't work. Both versions x86 and the x64 bit dont. Tested on Windows 2012 4-bit machine. 

     

    Okay updating this based on the totally insane response from Level 2 support. This probe has oppisite logic.

     

     

    So when using and configuring this the ntperf probe, you have to read the metric your setting up as the following statement: "The expected value is: X = Y, if not then trigger an alert" 

     

    For instnace you want to monitor Memory Usage and trigger an alert when Memory usage < 100MB. (just and example) you would setup your WMI query on Memory then Available Memory. On the "Alarm on Value" tab, you would think you would set the comparison to  < 100 trigger an alert. But no. The probes logic is I expect the value to be < 100. So if memory was 200 it would trigger an alert. You would have to switch the Threshold Operator to > 100. :smileyfrustrated:

     

    So this is going right back to development for them to fix this. No other probe follows this logic. We don't configure the CDM probe to say I don't expect the C:\ drive to be < 10% but if it does then throw an alert. No we just say, if Free Space < 10 then throw an alert. 



  • 26.  Re: **** Defect Announcements ****

    Posted 01-22-2015 06:45 PM

    @: That's basically what I fear most: Upgrade a probe and see no crash and no unnessary events but instead the probe will simply not generate events when expected. Not easy to detect as a customer, and I wonder how the described defect passed any kind of automated test framework



  • 27.  Re: **** Defect Announcements ****

    Posted 01-30-2015 12:50 AM

    In the list designer, if you have a list report, it doesn't populate an "Info" column if the type of the column is QoS or Origin.

     

    Backline has determined that this is actually a known defect with ID DE37741.


    The release date for a fix is unscheduled, as of this time.

     

    Thought we need a UI bug to mix things up with all the probe issues listed....

     

    -Garin



  • 28.  Re: **** Defect Announcements ****

    Posted 01-30-2015 01:10 AM

    If you want an UI bug I just found this one:

    UMP v7.5 SLM Portlet. If you click on the wrench icon and then do "Preferences" I finally found this wonderful feature that would of made finding QoS Metrics for a particular Client - Robot - Probe about 1000x easier. 

     

    If you were familiar with the olf SLM Windows GUI tool you could create new groupings and ORDER BY: ORGIN - ROBOT - PROBE. 

     

    Well in that wrench - preferences GUI, if you tired to change anything then saved, going back to the FULL SLM view broke the SLM Portlet completely. 

    L2 Backline response was response was Upgrade to 8.1. Oh thank you, thank you...



  • 29.  Re: **** Defect Announcements ****

    Posted 02-03-2015 10:23 PM

    When a key in a config file contains the slash char, raw configure in both adminconsole and infrastructure manager will break the key at the slashes and convert it into a section nest.  This behaviour has been a feature of IM for years and has been faithfully carried forward into adminconsole 8.1 raw configure interface for a consistent user experience!

     

     

    Before raw config:

     

    <config>

            /opt/nimbus/robot = woops!

    </config>

     

    After corruption with raw config:

     

    <config>

       <opt>

             <nimbus>

                   robot = woops!

             </nimbus>

       </opt>

    </config>

     

    Adding some/**bleep**/key via raw config is equally consistent.



  • 30.  Re: **** Defect Announcements ****

    Posted 04-01-2015 07:39 PM

    Or add fake report_engine key in data_engine raw config and makes the SLM rich client keeps working

     

    guillaume



  • 31.  Re: **** Defect Announcements ****

    Posted 02-03-2015 05:18 AM

    Found a defect with the CDM probe (v5.10) which has been acknowledged by support as a logged defect with ID# DE39700

     

    The problem is that when using the Admin Console to configure the probe to capture QOS data for the Disk Read B/S and Disk Write B/S metrics by selecting "Publish Data", it generates alarms even though you don't tick the "Publish Alarms" check box.

     

    Workaround from support is:

     

    1. In the Admin Console first check the “high and low” threshold(NOT PUBLISH ALARMS) for the same metric and then uncheck and click save.

     

    2. Or you can Open the cfg in notepad ,under the disk>>alarm section you will get these 6 section according to QOS enabled
    1. disktotal error
    2. disktotal warning
    3. diskread error
    4. diskread warning
    5. diskwrite error
    6. diskwrite warning
    In the above sections, all 6 or as per your need add a key active with value no “active = no” to disable the alarms.
    For Eq:
    <diskread error>
    message = DiskRateError
    description = Read Throughput
    threshold = 0
    active = no
    </diskread error> 

     

    The CDM probe v5.11 was released a day or two after I was given this defect number, so I doubt the defect has been resolved yet, and there is no mention of it in the release notes.



  • 32.  Re: **** Defect Announcements ****

    Posted 02-03-2015 05:26 AM

    There was a defect with hub 7.61 (the GA version that was released with UIM 8.0) which was performance related, and caused issues with being able to open the interface_traffic configuration (refer to this article for more info). Anyway, the issue was resolved in a release of hub 7.62 which wasn't published in the archive, and was only available from support. I was able to verify with support that the new hub version resolved my exact issue as per the article.

     

    However, after upgrading UIM to 8.1, and subsequently the hub probes being upgraded to the new GA version of 7.63, I began experiencing the same problem again. After logging another ticket with support, it was referred to backline, and I was told that the fix that was implemented in hub 7.62 was LEFT OUT of the 7.63 build!! A new hub release coming soon will apparently have the fix back in it again.

     



  • 33.  Re: **** Defect Announcements ****

    Posted 02-03-2015 10:46 PM

    Unfortunately this is another case of snatching defeat out of the jaws of victory.

     

    7.62 was a point release they pulled together for me after I ran into the same issue.  The bug fix didn't make it in time for the 7.63 release cutoff but should be in the next hub version.  I knew that at the time and I think even reported it earlier on this thread.  The back end discussion at the time was that they were moving toward more rigid release guidlines to keep core components from being so buggy.  

     

    Unfortunately their versioning and lack of an effective information sharing process undermines both the rapid response to the bug which was top notch in my experience, and their best effort to move toward a release methodology that promotes a more predictable and consistent piece of software, something we're all begging for.  

     

    If release notes had documented the known issues and reference the bugs in a bug tracker maybe even noting that hotfix 7.62 is not included in release 7.63, nobody would be shocked.  Hey, better yet, make sure that support can tag the hotfix with a usage note when handing it out.  The developers working on it let me know the upcoming release wouldn't have the hotfix and that info already passed through support.  They could have told you the same information when providing the hotfix and you wouldn't have wasted your time on an apparent bug reversion.

     

    So much was done right that they almost deserve a little inzone dance for good work and quick respons, but somehow they convert it into a terrible customer experience.



  • 34.  Re: **** Defect Announcements ****

    Posted 02-03-2015 11:15 PM

    Perl-SDK >= 5.05 in windows 32 bit segfaults.  5.04 works bug has a bug related to daemon probes and lacks CMDB/NIS2 related calls.

     

    Version 5.05 and 5.06 segfaults when loading the nimbus apis. This happens with both toolkit and Nimbus::API.

     

    5.04 works as expected. I tracked this down to the "bootstrap Nimbus::API $VERSION;" line which loads the API.dll. Leaving 5.05 or 5.06 deployed and just swapping in the 5.04 API.dll fixes the issue.  (Not a recommended workaround, just a conformation of where segfault is.)

     

    Public change logs are sparse in 5.05 and non-existent for 5.06. 5.05/6 works well in 64 bit windows and all variations of Unix.

     

    A workaround is to use SDK_Perl 5.04, but this lacks the NIS2 calls added in 5.06.  5.05 also contains a bug preventing Win32 from automatically registering the port when establishing a server session.

     

    The bug in question is listed as fixed defect in 5.05 "Win32 server port retrieval fix." I can tell you exactly what that means. When you start a Nimbus::smileyfrustrated:ession->server, it's supposed to report the port back to the controller. The end result is that server session is unusable, thus creating a daemon (not timed) probe with Perl_SDK 5.04 fails unless you use a workaround by to manually register the port with the controller after establishing the session.

     

    my $session = Nimbus::smileyfrustrated:ession->new($prgname);

    if ( $session->server(NIMPORT_ANY, \&timeout, \&restart) == NIME_OK ) {

         nimRegisterProbe($prgname,$ENV{NIMPROBEPORT});

    } else {

         die "sucks" ;

    }

     

    Of course this still won't allow you to associate alarms with ci data since those functions are not in 5.04.  Kind of makes you wonder how fixing the Win32 specific bug in 5.05 was tested if the dll causes segfaults?

     

    5.05 was released May 2013.

    5.06 was released Dec 2013.

    I discovered and reported the segfault Nov 2013.

    No ETA yet on a bug fix.



  • 35.  UIM-8.1 ppm-3.02 3.03 and 3.04 upgrades do not work.

    Posted 02-12-2015 01:01 AM

    The good news is that the 3.04 ppm fixes the infinite recursion ppm bug reported earlier.  The bad news is that you have to delete the probe from you robot, then install 3.04 fresh, but even that might not be 100% effective.

     

    I upgraded the probe and it wasn't fixed. Checked the log and it's reporting itself as 3.02. Note, I had upgraded to 3.03 previously, and the controller reported that 3.04 was installed.

    I found the problem in lib. Apparently there are versioned ppm.jar files being installed and the old ones are never cleaned up. I believe java is trying ppm.jar first when resolving the library.  I'm not sure how it chooses but it was consistently coming up with 3.02 which I believe is the default installed with UIM-8.1.   Anyway, moving all of the ppm*jar files except the 3.04 versioned one out of nimroot/probes/service/ppm/lib got it to successfully strart and report 3.04 which then successfully passed the bug test.

    Unfortunately this looks like another bug, and whomever is using 3.03 expecting those bugs to be fixed after an upgrade may not be seeing any of those patches working either. :smileysad:

    [nimroot/probe/service/ppm]$ ls -lt lib/ppm*.jar
    -rwxrwSrwx 1 nimbus sshusers 6340390 Feb 11 15:14 lib/ppm-3.0.4.jar
    -rwxrwSrwx 1 nimbus sshusers 6336302 Jan 7 15:06 lib/ppm-3.0.3.jar
    -rwxrwSrwx 1 nimbus sshusers 6336316 Jan 6 11:30 lib/ppm-3.0.2.jar
    -rwxrwSrwx 1 nimbus sshusers 6291487 Dec 19 15:39 lib/ppm.jar

     

    You might be able to work around this by uninstalling ppm first, but if you have already upgraded, you may have lost a reference to one of the versioned ppm.jar files and it wouldn't be uninstalled.

     

    MPSE seems to leave crufty old version libraries behind too after updates, but at least it calls the version mpse jar in startup args.  There is no simple way of knowing how many of the updated libraries it requires to operate as expected or which ones it's using.

     

    [nimroot/probes/service/mpse]$ ls -lt cimanager_bundle-8* ctdbase-2* ctdextension-2* ctdgraph-2* ctdtemplate-2* ctdutil-2* mpse-1* nimclient-core-1* nimclient-model-1* nimclient-probes-1* nimsoft-SDK-2* pf-common-2* stax-api-1*
    -rwxrwSrwx 1 nimbus sshusers 23346 Jan 6 11:28 stax-api-1.0-2.jar
    -rwxrwSrwx 1 nimbus sshusers 26514 Jan 6 11:28 stax-api-1.0.1.jar
    -rwxrwSrwx 1 nimbus sshusers 356419 Jan 6 11:28 pf-common-2.5.0.jar
    -rwxrwSrwx 1 nimbus sshusers 387639 Jan 6 11:28 mpse-1.6.4.jar
    -rwxrwSrwx 1 nimbus sshusers 26161 Jan 6 11:28 nimclient-core-1.6.5.jar
    -rwxrwSrwx 1 nimbus sshusers 6200 Jan 6 11:28 nimclient-model-1.6.5.jar
    -rwxrwSrwx 1 nimbus sshusers 25126 Jan 6 11:28 nimclient-probes-1.6.5.jar
    -rwxrwSrwx 1 nimbus sshusers 178299 Jan 6 11:28 nimsoft-SDK-2.9.25.jar
    -rwxrwSrwx 1 nimbus sshusers 337236 Jan 6 11:28 ctdbase-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 223457 Jan 6 11:28 ctdextension-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 65866 Jan 6 11:28 ctdgraph-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 16008 Jan 6 11:28 ctdtemplate-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 50872 Jan 6 11:28 ctdutil-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 1518021 Jan 6 11:28 cimanager_bundle-8.1.0-2014-12-18.jar
    -rwxrwSrwx 1 nimbus sshusers 292472 Jan 5 14:11 pf-common-2.4.0.jar
    -rwxrwSrwx 1 nimbus sshusers 176378 Jan 5 14:11 nimsoft-SDK-2.9.23.jar
    -rwxrwSrwx 1 nimbus sshusers 24915 Jan 5 14:11 nimclient-probes-1.6.4.jar
    -rwxrwSrwx 1 nimbus sshusers 6198 Jan 5 14:11 nimclient-model-1.6.4.jar
    -rwxrwSrwx 1 nimbus sshusers 26161 Jan 5 14:11 nimclient-core-1.6.4.jar
    -rwxrwSrwx 1 nimbus sshusers 387102 Jan 5 14:11 mpse-1.6.3.jar
    -rwxrwSrwx 1 nimbus sshusers 48460 Jan 5 14:11 ctdutil-2.3.0.jar
    -rwxrwSrwx 1 nimbus sshusers 15365 Jan 5 14:11 ctdtemplate-2.3.0.jar
    -rwxrwSrwx 1 nimbus sshusers 63258 Jan 5 14:11 ctdgraph-2.3.0.jar
    -rwxrwSrwx 1 nimbus sshusers 223453 Jan 5 14:11 ctdextension-2.3.0.jar
    -rwxrwSrwx 1 nimbus sshusers 310017 Jan 5 14:11 ctdbase-2.3.0.jar
    -rwxrwSrwx 1 nimbus sshusers 1468152 Jan 5 14:11 cimanager_bundle-8.0.0-2014-09-12.jar



  • 36.  Re: UIM-8.1 ppm-3.02 3.03 and 3.04 upgrades do not work.

    Posted 02-12-2015 01:26 AM

    Yep - your ppm install looks a lot like mine. Many files with the version numbers as part of the name. Many of which don't match the reported installed version.

     

    I had raised the 3.02/3.03 version number thing with support at one point and was told that it was probably just not updated in the new version. It loading the wrong thing though would certainly explain some of the behavior I had seen.

     

    Thankfully with 8.1 I don't have to worry what version of ppm is installed or where as without anything that uses ppm working there's nothing that cares - Grin,

     

    -Garin 



  • 37.  UMP 8.1 lucene reindex doesn't replicate through UMP nodes.

    Posted 02-17-2015 12:03 PM

    Problem:

    When a user logs into UMP for the first time in a multinode UMP-environment, the user only appears under that node's "Users and Organizations" in Control anel. Apparently this is a Liferay issue.

     

    Workaround:

    In Control Panel > Server Administration do "Reindex all search indexes" on nodes that are missing users.



  • 38.  UMP 8.1 USM "interfaces" tab doesn't display metrics if value is below 1

    Posted 02-19-2015 01:59 PM

    Discards, Errors and Utilization columns don't display metrics if the most recent value is < 1. This has been recognized by CA and raised as a defect.

     

    -jon



  • 39.  cdm 5.21 pulled from the archive

    Posted 02-24-2015 11:06 AM

    for a short while cdm 5.21 was GA and available in the archive. Now it is back to 5.11 so in case you are already using cdm >5.11 you may want to downgrade until issues are resolved (whatever these issues are)



  • 40.  Re: cdm 5.21 pulled from the archive

    Posted 02-24-2015 07:44 PM
    i was informed by support that the issues about cdm 5.21 were at documentation level only


  • 41.  usertags dont get updated on existing alarms

    Posted 02-26-2015 04:55 PM