Automic Workload Automation

  • Private Community
Back to discussions
Expand all | Collapse all

Notes on Upgrade from V9 sp8 to V11.2

  • 1.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 19, 2015 05:28 PM
    Hi,

    For those of you at the Automic World, you probably heard that I was going to test the upgrade process from V9 SP8 to V11.2. I wanted to give some updates.

    Summary - Once I got the different pieces of software and OS's working, the upgrade process is quite smooth. It took 8.5 hours, which is way too long for us to have as downtime, but can probably be reduced. Looking at the logs we can tell that most of that time is attributed to several tables, that although we archive, might still be too big.

    Issues faced:
    1. X-Windows is required so that you get a graphical interface for both the Automic Utilities and Oracle. It turns out attempting to run X via mac or pc over the network is PAINFULLY slow. We realized that the Automic utilities do not have to be run from the same machine as the automation engine, so we brought up a virtual-box vm with CentOS and gnome that made it much easier and faster from the X side of things.
    2. Full Oracle client is required. We can no longer use the Instant client. I attempted to install the Full Oracle client on ubuntu which is the preferred linux distro for us. It turns out that Oracle does not support ubuntu, and the amount of effort and sub packages required to get the full client installed on ubuntu was too much. So we brought up a CentOS instance where we could install the full oracle client. This CentOS instance also has the Automation Engine binaries on it. One of the issues was remembering to set the Unix variable NLS_LANG for the users running the Oracle Client. This has to match the settings on our Oracle DB. This variable setting is also required for the user that runs the Automation Engine (which in unix is root). Otherwise, the service manager is unable to start the WPs and CPs.

    The Virtual-Box VM with CentOS allowed us to run the Automic DB Load utility. After several retries due to Oracle and Automic settings, the process went through with no blocking errors. The Load utility running the Upgrade script for 11.2 checks if your DB is on 8, 9, 10, 11, or even 11.1, and automatically takes the system through each of the steps. Which was very nice, as from the documentation, it seemed that I would have to do each step individually.

    Like I said in the summary, the entire Load took 8.5 hours. The parts that took the longest were
    During the Database changes, an AV table, that had 198 million records, was insert into av_new (select * from av). This took 1.5 hours.
    During the Intial Load steps, the MSG and MSGTX tables, which don't exist in v9, took 40 minutes and 1.5 hours respectively. Then the OT table took 2.5 hours.

    If the DB is empty, I can easily see it taking the 30 minutes as was tested by Automic. But as that is not the case with most people, I would recommend figuring out exactly how much and what tables to archive before starting this process.

    I'm now trying to get the User Interface working, but I'm getting an error there as well. I'll update once I get that working.

    Good luck on all your projects.
    Samah


  • 2.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 19, 2015 06:46 PM
    Thank you SamahSohrab607228  for starting a V11.2 post!

    We are in the process of migrating from V9.SP5 to V11.2.  We are running two instances (DEV with 5 clients, and PROD with 1 client) on Windows and SQLServer.  We also need to upgrade the WindowsOS, Java, and SQLServer at the same time.  Our environment is relatively simple with only 5 remote agents.

    We enlisted the help of an Automic professional.  In a 3 day session he helped my team bring up 100% of the new DEV system and completed about 80% of PROD.  He is contracted to come back when we are close to going live.  Our approach (at his recommendation) has been to stand up our two brand new servers, install V11.2 from scratch onto them, and export-import all objects to minimize downtime.  On go-live date we will rename the servers and make sure the server that runs RA/FTP is assigned the desired IP address.  When we go live we will move our V9 history into the archives and then shut it down.

    We are running both V9 and V11 remote agents by virtue of using different ports.

    V11.2 Defect found;  Any time we ran a job on a UNIX agent that included a :REGISTER_OUTFILE function we would get a hard failure.  Support discovered there is a missing character in object REGISTER_OUTFILE.UNIX under client 0000.  There is a line of code with &UC_JOBMD that was deployed as &C_JOBMD.  We added the "U" to fix it.

    I'll be watching this thread for other's V11.2 experiences!


  • 3.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 19, 2015 07:52 PM
    Thanks a lot for bringing your Lessons Learnt to the Forum.  Will bookmark this Topic, so we will use it when we start the planning for upgrade v10 to v11 (some issues could still be similar)


  • 4.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 19, 2015 08:27 PM
    Pete,

    I was thinking of using that approach. But the Import Export take so long. Also I lose all statistical history. And the second one I want.


  • 5.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 20, 2015 04:31 AM
    Hi,

    Thanks for sharing, Samah.
    We are currently on AE V9 SP11 Hotfix 1, the Automation Engines are running on virtual Windows servers (VMWare) and the database is Oracle 11.2.
    Sometime next year we will upgrade to V11.2+, and it's always nice to hear from someone who has already upgraded.

    Best regards
    Keld.


  • 6.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 20, 2015 10:37 AM
    A HUGE thanks for taking the time to write this up SamahSohrab607228 !! We'll be working toward a similar upgrade path from v9 sp11 to 11.2, so your experience will no doubt help us tremendously.


  • 7.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 20, 2015 11:18 AM
    "X-Windows is required so that you get a graphical interface for both the Automic Utilities and Oracle." 

    Speaking from a v8 -> v10 upgrade experience, I was able to use the Windows Utilities (    Utility\windows\x64 in the image) from my Win8 PC to do a load on our Linux + Oracle system. Functionality was equivalent to running from X-Windows utilities on a different server.

    The only "tricky" part was adjusting the config files and making sure I had the right Java and 64 bit Oracle client installed. Also, I have super low latency to my env, so local works well.


  • 8.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 20, 2015 07:37 PM
    A new update. It turns out that the WP will not start on this upgraded system. I'm dealing with Support in trying to figure out why. The CP starts just fine.


  • 9.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 22, 2015 06:54 PM
    Actually there is no need to install X Windows or do kinds of workarounds to get the GUI for the Utilities running. Fine to just start via command line. 
    Anyone done a V8 to V11 upgrade?


  • 10.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 23, 2015 06:31 PM
    It seems that when I pass my V9 Reporting Tool XML files to the V11 version of UCYBDBRT, it does not produce the expected results.  It seems the quickest solution is to rebuild them with the V11 Reporting Tool.  Thankfully I have less than 10 reports to rebuild.

    UPDATE:
    I'm finding that usually I can just edit the xml and upgrade the "Query version" attribute to get the desired results without changing anything else.  Too bad that UCYBDBRT isn't smart enough to accept V9 specifications.


  • 11.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 25, 2015 05:29 PM
    I've discovered a behavior change with the get_msg_txt function.  Under V9 if I passed it a text string for the first argument it would give me a null result and everything would go happily on its way.  But under V11 it causes the entire script to fail with U00020440 "Non numeric input not allowed".

    This behavior change breaks my solutions that follow a suggestion that I posted here;
    https://community.automic.com/discussion/comment/19897#Comment_19897
    I suggested putting the word "failed" inside of &UC_CAUSE_STATE which is later passed to the get_msg_txt.  So my workaround is to replace the word "failed" with a meaningless numeric value.



  • 12.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 25, 2015 07:12 PM
    I've discovered a behavior change with the construction of JOBP objects.  

    One typically starts construction from the template object called JOBP that is stored in client 0000.  In our case the END object in this template was set to "else skip" on the dependencies tab, but this was always overridden with "else block" when I drew a dependency into the END object based upon my personal "workflow" settings.

    Under V11.2 however, I've discovered that drawing a dependency into the END object does not override the "else skip".

    My solution is to modify the template to "else block", which is a much safer default in the first place.


  • 13.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 27, 2015 08:43 AM
    Hi,

    I suppose it's still required to do an update from V9 to V10 before updating to V11, meaning that it's not possible to go directly from V9 to V11, right ??

    Regards.


  • 14.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 30, 2015 11:52 AM
    I can't answer Keld_Mollnitzs  question because we did not convert our database.  But I think I've heard that they do have the ability to convert a V9 database directly to a V11.2 database now?


  • 15.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 30, 2015 03:53 PM
    >@</p>Keld Mollnitz"  The dbload tool with the 11.2 upgrade sql will automatically take you from v9 to v11.2. Internally it goes through each step from 9->10->11->11.1->11.2. But you won't need to worry about that.


  • 16.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Nov 30, 2015 04:37 PM
    We discovered a V9.SP5-to-V11.2 promptset behavior change.  In our V9.SP5 we didn't use the "Quotes" feature on promptset text fields very often.  It always works fine with a double quote.  But I now understand that it never did work with a single quote for us.

    We found several instances in production where someone had coded a single quote into this feature, but UC4 basically ignored it.  When this code was imported into V11.2 it caused failures because the single quote was applied when we did not want it to be applied.

    Here is the V9 SQLServer query that found the cases that needed to be corrected (we removed the single quote rule from the promptset object);

    SELECT oh_name, opse_name, opse_quotes
      FROM UC4.dbo.OPSA, uc4.dbo.oh, uc4.dbo.opse 
    where opsa_oh_idnr = oh_idnr
    and   opse_oh_idnr = oh_idnr
    and opse_quotes = ''''
    order by 1, 2


  • 17.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Dec 01, 2015 12:11 PM

    There is a nice video walkthrough of performing an upgrade from version 11.2 to newer versions (must get to 11.2 first!). We were wondering what this will look like, so thought I'd pass along:

     

    Other videos posted here: http://automic.com/resources/video



  • 18.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Dec 02, 2015 11:18 AM

    The Load utility running the Upgrade script for 11.2 checks if your DB is on 8, 9, 10, 11, or even 11.1, and automatically takes the system through each of the steps. Which was very nice, as from the documentation, it seemed that I would have to do each step individually.
    In a discussion forum post entitled upgrade scenario to V11.1, posted on 9 April 2015, FrankMuffke stated that “older versions must be updated to V10 first, then to V11.1.”
    Perhaps this restriction applied only to v11.1. The v9→V11.2 upgrade instructions certainly make no mention of a requirement to upgrade the database in multiple steps.


  • 19.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Dec 03, 2015 03:43 PM
    I talked to some of the developer previously and they have confirmed that upgrading from V9 -->V11.2  can be upgraded directly and does not need to go through V10.


  • 20.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 11, 2016 12:24 PM

    Has anybody experienced any issues with a V9 agent running in a newly upgraded V11 Engine?  Normally you can run with an older agent when you upgrade from 1 version to the next, but If I go from V9 to V11 I'm skipping V10. Has anybody ran a V9 agent in V11? thanks



  • 21.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 21, 2016 12:41 PM
    RonGates608562 
    We are upgrading all agents to V11, so no, I haven't tested V9 agents against V11.


  • 22.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 21, 2016 12:43 PM
    We have run into a showstopper for our upgrade to 11.2.  The autoforecast() function is not working.  It works fine when there is hardly anything scheduled, but as soon as I introduce our most complex workflow into the schedule mix, the function hangs.  I've been working with support for over a month on this.  They have been able to recreate the problem from my transport case files, but we don't have any answers from development regarding a cause.  (We run on Windows and SQLServer.)

    Has anyone else tested autoforecast in 11.2, and what have your results been?  I'm hoping to find another datacenter with the same problem to help raise the priority of this with Automic?  (Support says there has also been some performance issues that they are actively working on solving.)


  • 23.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 21, 2016 02:08 PM
    I hope to have my 11.2 system up today / tomorrow and I can check for you.  I'd be interested to know more specifics about the performance issues though.


  • 24.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 21, 2016 04:11 PM
    If I understood him correctly, the performance issues were only with the autoforecast() function.


  • 25.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 21, 2016 05:05 PM
    We’ve seen a problem with 11.2 that is a potential showstopper for us too. When the problem happens, a USER object gets corrupted and the user cannot log in. Specifically, the LDAP distinguished name of the USER objects gets overwritten with the DN of a different user. The DN inserted is usually the ID of a UC4 administrator. We haven’t been able to determine what causes the problem, but we did devise a way to identify affected USER objects.

    The following query will list USER objects whose DN and ID do not match.
    select USR.USR_OH_Idnr,USR.USR_FirstName,USR.USR_LastName,OH.OH_Name,USR.USR_DstngshdName
    from USR,OH
    where USR.USR_OH_Idnr = OH.OH_Idnr
    and SUBSTR(OH.OH_Name,1,6) <> SUBSTR(USR.USR_DstngshdName,4,6)
    (Note that our user IDs are all 6 digits long.)

    When we have identified a corrupted USER object, the fix is simple enough: an admin just has to open the USER object and click the button to synchronize the information from the LDAP server.

    Has anyone else seen this problem?


  • 26.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 29, 2016 05:59 AM
    RonGates608562 
    We are on AE v9 SP11 Hotfix 1 and planning on going directly to 11.2. Most of our agents are on v9 but we even have a lot of Unix (Solaris) Agents on v8 (!)
    I asked Support and this is the answer:

    Basically Automic grants downward compatibility Automation Engine > Agent from one Major Version (that means V8 Agents can run on a V9 Automation Engine within the features and possibilites of Version 8)

    Version 11 is 3 levels above Version 8.00A, so the compatibility cannot be granted.

    Technically it might be possible to run the Agents of this Version with a Version 11 system, but we strongly recommend to do a careful testing in this constillation on the Customer Environment, AND we also strongly recommend to update the Agents as soon as possible to the same major Version as the Automation Engine has.

    If the tests are not successful, the Agents have to be updated in same timely manner as the Automation Engine. 


  • 27.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jan 29, 2016 08:58 AM
    Michael_Lowry - just wondering how you identify a corrupted user object?  Just configured LDAP here and so far no issues, but just wondering what the symptoms of this would be.  Users can't login?


  • 28.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Feb 03, 2016 10:09 AM
    Originally, we identified corrupted user objects only based on complaints from users who could not log in. Now we have a more structured way to identify the affected objects. The following SQL query will list the USER objects in which the first six characters of the user ID does not match the relevant portion of the LDAP distinguished name.
    select OH.OH_Client, USR.USR_OH_Idnr, USR.USR_FirstName, USR.USR_LastName,
    OH.OH_Name, USR.USR_DstngshdName
    from USR,OH
    where USR.USR_OH_Idnr = OH.OH_Idnr
    and USR.USR_LDAP = 1
    and OH.OH_Client = 1
    and SUBSTR(OH.OH_Name,1,6) <> SUBSTR(USR.USR_DstngshdName,4,6)
    Note: our user IDs are all 6 characters long. If your IDs are different, you may need to adjust the SUBSTR bits.

    By the way, Automic has opened PRB00113183 for this problem. Feel free to subscribe to this problem ticket.


  • 29.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Feb 03, 2016 10:22 AM
    Thanks Michael_Lowry .  Tried subscribing to that problem, but can't find it - it won't come up when I do a search.  ???

    There should be an easy way to see current / known issues with the release you are on.  That would be helpful to see what issues are out there that you haven't run into yet (but could).  Is there a way to do this?


  • 30.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Feb 03, 2016 10:28 AM
    @"Laura Albrecht" : I do not know how Automic plans to handle subscriptions to problem tickets that were opened by other customers. Presumably, they will want to be sure that they do not reveal any information about one  customer to other customers. It does make a lot of sense though, to be able to see a list of current known issues, and to be able to subscribe to the ones that are most of interest. Perhaps problem tickets will by anonymized, or perhaps only some data will be made publicly visible. Let’s see what Automic comes up with.

    I suggest you open an information request and ask for how to subscribe to this problem ticket.


  • 31.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Feb 26, 2016 06:53 PM
    On January 21 above, I posted that we had encountered a show-stopper with the AUTOFORECAST() function not working in V11.2.

    Development re-created the problem and discovered it was caused by File Transfer objects that were using host groups.  The workaround is to hard-code them to specific hosts.  They also said they are going to develop a fix.

    In the mean time, I'm happy with the workaround and am going to proceed with planning for our upgrade to V11.2.  However V11.2.1 is supposed to come out soon, and I will probably apply this upgrade before proceeding.

    Pete


  • 32.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Feb 27, 2016 04:24 AM
    We have opted to postpone our long-planned upgrade from V9 to V11 to allow more time for testing. Barring unforeseen problems, we too will upgrade straight to 11.2.1, slated for release in a few days.


  • 33.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 04, 2016 08:27 PM
    I implemented 11.2.1 in our non-productions today.  When the developers started banging away, we discovered that the new security setting "FileEvents: Start without Login object specified" must be turned on if you want to execute a workflow precondition that checks the existence of a file in your file system.  As is often the case, the error message was not very helpful in steering us towards the solution.   (which is why I thought this was worth posting.)   

    Error message:
    U00000009 '': Access denied




  • 34.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 05, 2016 03:52 AM
    ... we discovered that the new security setting "FileEvents: Start without Login object specified" must be turned on if you want to execute a workflow precondition that checks the existence of a file in your file system.
    I wonder if file checks (CHECK FILE) in task pre-conditions stop working if ANONYMOUS_FT is set to N in UC_HOSTCHAR_DEFAULT.


  • 35.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 14, 2016 02:48 PM

    I am doing my first trial run in an isolated environment migrating from V9SP11HF1 to V11.2. in the AIX environment running Oracle DB.

    The db upgrade is taking a LONG time.. .I understand that its a big jump from V9 to V11.. .but does anybody have any timings on how long the db update step takes?  Any issues with that step?  any known performance enhancements that we can do to speed up that step?  It's been running now 3 hours.. I had the DBA validate and it's actually still working.. just taking a LONG time.. thanks for any input!!



  • 36.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 15, 2016 10:41 AM

     @Michael Lowry   I ran in to the LDAP id not working situation after upgrading from V9 to V11.2 

    I  couldn't get logged in with an LDAP id..

    You said to open the USER object and click synchronize... but how do you do that if you can't get logged in as an LDAP user?  I logged on as a non LDAP user and when I tried to sync ldap it said a non LDAP id can't sync a LDAP id.. So in your case it sounded like not all LDAP id's were effected?



  • 37.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 15, 2016 12:16 PM
    The bug I mentioned does not prevent LDAP synchronization. It causes USER objects to get corrupted (the DN gets overwritten with the DN of another user, from a previous sync.) The bug manifests itself only if the LDAP version is set to 2. Setting the LDAP version to 1 (VERSION in the LDAP Connection VARA) makes the problem go away. In case you’re interested, the problem ticket is PRB00113183, the Jira ticket is AE-8530, and it’s supposed to be fixed in 11.2.2.


  • 38.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 16, 2016 02:37 PM
    We upgraded our production environment from V9.SP5 to V11.2.1 today.  So far I've only had one surprise. We had a single application with a promptset value override that was coded with single-tick quotes around a string, and in V9 these single-tick quotes disappeared at run time and were not applied to the command line that uses the value.  When it ran in V11 these single-tick quotes were retained resulting in an application parameter error.   We scanned our database and found all parameter overrides that start with a single-tick and removed the ticks.

    I suspect the V9 behavior was a bug that has been fixed.


  • 39.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 17, 2016 12:50 PM
    ...we've been bit by a behavioral change that caused some of our Sunday morning tasks to not run;

    In V9.SP5 if you close a queue that is running a schedule, that schedule continues to prepare and submit tasks.

    In V11 if you close a queue that is running a schedule, that schedule object will not submit tasks.  Instead, they are given a status of "ENDED_TIMEOUT - Start time exceeded."

    I have an open ticket with support to find out more about this change of behavior.  I still don't know if we were depending upon a V9 bug, or this is a V11 bug, or what.

    UPDATE:  
    It has been determined that we were depending upon a V9.SP5 bug that has been corrected.


  • 40.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 20, 2016 11:58 AM
    I didn't want to post these until we were sure it wasn't just something we had coded funky, but we are dealing with two 11.2.1 bugs right now that are impacting our production operations. 

    INC00116917 - When a NOTIFY object tries to attach a registered external report to an email and that registered external report lives on an HPUX agent, it fails with error U0050031.  (We don't have an action plan on this one yet.)

    INC00116904 - When you register an external report on a WINDOWS agent and you also have a "WAIT" in the post-process script, UC4 loses the proper credentials for accessing that registered report and you can't open it from the UC4 client.   I've been told this should be fixed in 11.2.2 to be released June 30th 2016.

    Pete


  • 41.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 20, 2016 06:27 PM
    Last night we upgraded our system from 9 to 11.2.1 HF1. PLEASE DON"T DO IT!!!!!

    The system has several major bugs.
    1. GetMonitor/RefreshMonitor causes the DWP to go to a load of 100%, causing the ENTIRE system to be unresponsive to clients. Users can't login, those who are logged in, the system freeze. The only solution currently is to kill that DWP.
    2. Right click in Monitor window doesn't work on Unix (mac) user interface
    3. PWP crashes due to deadlock in DB after trying to access HSTA

    AVOID v 11.2.1 HF1


  • 42.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 21, 2016 04:11 PM
    I think you should be careful with blanket statements like this.  I've been on that version since it was released and had no issues.  Could be my very low usage of the product at this point - it is a new system, but there are usually many variables involved in bugs like this.  Platform of the AE's, db type, etc.  Just my 2 cents.  I'm on Solaris and Oracle BTW.


  • 43.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 21, 2016 04:28 PM

    Couldn't agree more that there are too many environmental differences between each of our datacenters to make blanket statements.  Our 11.2.1(Build.412) is going to be fine once I code a couple more workarounds, but that seems to be par for the course for any upgrade I've done.  We are on Windows2012 and SQLServer2014.

    Samah told me he is on 11.2.1(Build.414).



  • 44.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 26, 2016 01:10 PM
    Last night we upgraded our system from 9 to 11.2.1 HF1. PLEASE DON"T DO IT!!!!!

    The system has several major bugs.
    1. GetMonitor/RefreshMonitor causes the DWP to go to a load of 100%, causing the ENTIRE system to be unresponsive to clients. Users can't login, those who are logged in, the system freeze. The only solution currently is to kill that DWP.
    2. Right click in Monitor window doesn't work on Unix (mac) user interface
    3. PWP crashes due to deadlock in DB after trying to access HSTA

    AVOID v 11.2.1 HF1                       

    Hi Samah,

    I had taken a look at the logs that was sent in to support on the 20th.  From the logs it showing you are on version '11.2.1+build.412'SamahSohrab607228  - As for the 11.2.1 HF1 it should be '11.2.1+hf1.Build.440'  and was officially released yesterday (4/25/2016).

    Within the Hotfix that was released on 4/25/2016, it appear that some of the fixed within it would address the issue you are describing in regards to the DWP/PWP in your circumstance.   I know you are currently working with Automic Support & Development to finalize the upgrade.  Please update us on your upgrade to 11.2 once everything has finalized.

    @Laura & Pete, I would with agree with both of you on this, as blanket statement would not fit everyone as there would be too many environmental difference for everyone.  In this particular case, 11.2.1 had worked for you both, while Samah had encounter some issue unique to his environment.

    In this regards, if anyone is thinking of upgrading to the 11.2.1 or 11.2.1 HF1 (that was released on 4/25/2016) from an older version, I would recommend going through test environment prior and should you run into any issue, to contact Automic Support for the incident.

    Edit: I spoken to one our the Support Technician and they have confirm with Samah that the version he is actually referring to is AE v11.2.1  and not the Hotfix one.


  • 45.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 27, 2016 03:29 AM
    Luu Le said:
    I had taken a look at the logs that was sent in to support on the 20th.  From the logs it showing you are on version '11.2.1+build.412'  - As for the 11.2.1 HF1 it should be '11.2.1+hf1.Build.440'  and was officially released yesterday (4/25/2016).

    ...

    Edit: I spoken to one our the Support Technician and they have confirm with Samah that the version he is actually referring to is AE v11.2.1  and not the Hotfix one.
    Thanks for the clarification, Luu. My guess is that Samah mistook the service pack number for a hot fix number.

    This misunderstanding illustrates the importance of having a unified and consistent approach to versioning.


  • 46.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 27, 2016 07:53 PM
    We've implemented effective workarounds for all of our 11.2.1 issues, so feeling pretty good about it right now.

    Wanted to mention that the problem we found with using a WAIT in a post-process script also happens if you use a PREP_PROCESS_FILE function in the post-process script.  I was able to determine that this defect did not exist in 11.2, but it does exist in 11.2.1, and they hope to have it fixed in 11.2.2.  (It clobbers the credentials associated with registered reports so the registered reports can not be opened or attached to notify/mail objects.)

    Support could not re-create our HPUX registered report attachment issue so there must be something unique about our environment.  But we have coded around this issue as well so I'm not going to pursue it any further.

    Overall, V11 feels more stable than V9 to me.   And the upgrade process from V11.2 to V11.2.1 using the zero-down-time method was not too bad.


  • 47.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 28, 2016 08:03 AM
    Hi everyone,

    We are currently on V9 SP12 HF3 (AE is on Windows and DB is Oracle 11) and my plan is to upgrade to V11.x before July.

    Should I go for 11.2.1 HF1 (released 4/25/2016) or should I wait for 11.2.2 which according to petwir should be released June 30th 2016. ?

    Can anyone give some input/good advice on this??




  • 48.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 28, 2016 12:03 PM

    Well, I was shocked to find out that upgrading from 11.2 to 11.2.1 introduced a new bug for us.  The common assumption is that a newer version should not introduce new bugs, but they can.  The moral of the story is that no matter what version you go with, make sure you have a thorough enough test plan to exercise all of the product features that your datacenter depends upon.




  • 49.  Notes on Upgrade from V9 sp8 to V11.2

    Posted May 04, 2016 08:51 PM
    I need to apologize for my previous rant. We upgraded to 11.2.1 and had many issues. Bugs that we were not aware of and bugs that crashed the entire system. Our Automic system is part of our critical path, and we had scheduled a downtime that was extended because of these issues. So it became quite stressful.

    Automic has been very helpful in fixing the issues we had and responding quite quickly.
    We got them to release 11.2.1 HF 1 to fix the issue where Monitor of a workflow took over 2 minutes rather than 9 seconds, and refresh of the monitor would cause the WPs to get stuck at 100% usage for a long period of time.
    Then a week later they released 11.2.1 HF 2 to fix the issue where in the Unix client did not allow for Right Click in the Monitor view.
    Both of these were major issues for us, and it was quite frustrating during a major upgrade.
    There is still a bug where our Automatic Agent Assignment (HSTA) is not working, but it is a bug they are working on.

    Automic both from the Customer Service and Engineering has been very helpful and getting our issues resolved. Also the Zero Downtime Upgrade for the Hotfixes worked as well.

    We a generally happy with 11.2.1 HF2


  • 50.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 17, 2016 07:33 PM
    We were bit by another V9-2-V11.2.1 upgrade issue today that is unique to TransportCase migrations.

    We did not convert the database.  Instead we stood up a new system and used TransportCase to bring our objects across.   All of our LOGIN objects were working as expected until one of them required a change.  Staff made a minor modification to an important LOGIN object, hit "SAVE", and most of the rows disappeared.  He failed to see them disappear because you have to close-and-reopen the object to see what actually happened.  Then in the middle of the night things started to fail for lack of credentials.

    The solution was to pause operations and manually rebuild the contents of the LOGIN object.  It wasn't technically hard to repair, but one sure does dislike the high visibility of such a failure.  Thank goodness we have good password backups.  (We use the "Password Safe" utility as a password backup facility.)

    I've re-created the problem and sent it to support so that they will be aware.


  • 51.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 18, 2016 03:22 AM
    Pete Wirfs wrote:
     
    We were bit by another V9-2-V11.2.1 upgrade issue today that is unique to TransportCase migrations.

    We did not convert the database.  Instead we stood up a new system and used TransportCase to bring our objects across.   All of our LOGIN objects were working as expected until one of them required a change.  Staff made a minor modification to an important LOGIN object, hit "SAVE", and most of the rows disappeared.  He failed to see them disappear because you have to close-and-reopen the object to see what actually happened.  Then in the middle of the night things started to fail for lack of credentials.
    Does the bug affect only LOGIN objects that have been loaded via the TrasportCase mechanism? Or does it also affect, say, LOGIN objects newly created in v11.2.1? Thanks.


  • 52.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 20, 2016 12:17 PM
    I just experimented with building a new LOGIN object and could *NOT* recreate the problem.   And our transported objects seem fine once they have been repaired.

    I'm curious to find out what support has to say about this.  I'm suspicious that it might be related to having our own custom "TYPE"s in the system(?)


  • 53.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 20, 2016 12:29 PM
    Pete Wirfs wrote:
     
    I just experimented with building a new LOGIN object and could recreate the problem.   And our transported objects seem fine once they have been repaired.

    I'm curious to find out what support has to say about this.  I'm suspicious that it might be related to having our own custom "TYPE"s in the system(?)
    I assume you’re referring to UC_LOGIN_TYPES. Were only the entries associated with custom login types affected?

    By the way, we found a strange bug related to this feature: GET_LOGIN does not work properly if the LOGIN TYPE in UC_LOGIN_TYPES has a & character in any value column. (This problem is documented in PRB00117767.)


  • 54.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 20, 2016 12:40 PM
    The one that hit us hard utilized only one TYPE that is in UC_LOGIN_TYPES.   It was where we store server credentials for PowerShell solutions that connect to remote windows servers.


  • 55.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 20, 2016 12:47 PM
    IMPORTANT CORRECTION!  One of my posts above said "I COULD RECREATE THE PROBLEM" from building a new object.  This was a very misleading typo!   It was supposed to read "I COULD *NOT* RECREATE THE PROBLEM!".  I'm correcting that post now... (Dang monday morning anyway!)


  • 56.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Jun 28, 2016 08:05 AM
    Michael Lowry said:
    ... we discovered that the new security setting "FileEvents: Start without Login object specified" must be turned on if you want to execute a workflow precondition that checks the existence of a file in your file system.
    I wonder if file checks (CHECK FILE) in task pre-conditions stop working if ANONYMOUS_FT is set to N in UC_HOSTCHAR_DEFAULT.
    I just tested this. File checks in task pre-conditions work fine even if ANONYMOUS_FT is set to N in UC_HOSTCHAR_DEFAULT.


  • 57.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Aug 10, 2016 11:22 AM
    petwir I am trying to see which approach will work best for our environment. The documentation that I see (http://docs.automic.com/documentation/AE/11.2/english/AE_WEBHELP/help.htm#ucsadd.htm) is for duplicating and updating the db and the core components. Do you have any documentation for the approach you used? Or can I just use the "new installation documentation" and then use transport case to move the jobs?
    PS: We are still on V9sp8


  • 58.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Aug 13, 2016 02:07 AM
    JGi604607  I'm on vacation for the next couple of weeks so I won't be readily available to answer your questions,  

    We did a new installation on new servers and used transportcase.  This made it easy for us to also upgrade our server OS and our SQLServer at the same time.  Can't say we followed any specific documentation though.  We figured out what would work best for us.

    Using this method had both pros and cons.  We enjoyed being able to test V9 and V11 in parallel, as we also fired up parallel agents.  On cutover day, there was very little downtime because we literally just turned one system off and the other system on.  However this also caused us to lose our run history so we dumped it into the archive system.


  • 59.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Aug 15, 2016 07:51 AM
    I find this thread very interesting.  Is this how most customers are doing this?  Upgrading in parallel?  I have always installed in place - of course taking proper precautions prior so that an easy backout could occur.  Never had any issues.  We did prior training to the ugprade so the users weren't shocked when the upgrade happened and especially now with ZDU - I don't see any reason to do an upgrade in parallel like this.  Is the main reason the downtime?  Or learning curve?  Just curious.


  • 60.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Aug 15, 2016 08:21 AM
    In most of the case the reason to do a migration using transport case is the downtime specially if you have very large DB with tons of reports and statistics. Even if you need to keep history for legal reasons only, the quickest way is to drop all of it and do a definition only migration. Archive all history later for legal can take hours or days on the "old" database, it will not impact live production in that case.

    I would suggest the following to do it simply (more or less of course).

    - install your new configuration in 11.x version but do not initialize DB
    - create a separate DB for 11.x instance and initialise it with the same version as the current one you want to migrate.
    - transport all you client 0000 (search all objet + transport) into the new DB client 0000 if you made many customization on objets or have your calendars etc in client 0000
    - use the client copy utility (ucybdbcc program) to copy all definitions of all clients from original DB to new DB. This is rather fast and allows you to check inconsitencies in the original DB.
    - once done and ok, run the upgrade with the ucybdbld program 11.x on the new DB. Must be also rather short compare to a full DB upgrade.
    - activate the 11.x server on the new DB and check results.

    Repeat process, after dropping all tables in 11.x version, until you are satisfied with the result.

    Note that in client copy you can keep the versions of objects (optional) but be careful with Login Objects and User Objects, password can be reset during copy process.


  • 61.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Aug 29, 2016 10:23 AM

    Interesting topics,

    Many thanks for your experience sharing.

    We also will migrate from our current V9SP12 to V11.2.



  • 62.  Notes on Upgrade from V9 sp8 to V11.2

    Posted Apr 03, 2017 04:53 PM
    SamahSohrab607228
    we are having the same problem.  CP's start fine.  WP's wont start after the upgrade from V9 to V11.
    how did you resolve this issue?