Forum Discussion

zubkoff_s's avatar
zubkoff_s
Level 4
11 years ago

Discovery Accelerator cannot search within ZIP files. Evidence. Documentation.

Hello, team.

 

Is a fact, that Discovery Accelerator cannot search within ZIP files. I've made several test cases and haven't found items with content inside ZIP file.

In this case EV indexes only file name and extension in ZIP file.

Can you please provide me documentation, where I can find this is really impossible to find items within ZIP file, and this isn't my fault and system misconfiguration. I have to provide this evidence for my boss.

 

Thanks.

  • Production team has confirmed the same system behavior in production environment: they didn't find the messages in DA, based on some words in ZIP files. 

    So, my test infrastructure, home lab and production environment have absolutely different configuration settings, version, hot fixes, patches, BUT have the same result with searches in ZIP files. 

    It doesn't look like coincidence. At least I haven't got clear opposite evidence from community.  

    Thanks for comment. 

17 Replies

Replies have been turned off for this discussion
  • Hi,

    Both DA Server should have same version and also should have installed EV binaries Version : EV 9 SP2

    9.0.2.1061.

    Following Path Should present on both DA Server :

    HEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\KVS\Enterprise Vault\Install

    If it not present on 1 DA Server please follow following steps on that DA Server :

    Steps to carry on Discovery Accelerator Server :

     

    1. Stop the “Enterprise Vault Discovery Accelerator” Service on Discovery Accelerator Server.
    2. Then Disable the “Enterprise Vault Discovery Accelerator” Service Startup Type.
    3. Reboot the Discovery Accelerator Server.
    4. Then install the correct EV binaries (9.0.2) on Discovery Accelerator Server but Don't Configure EV / Run Configuration Wizard otherwise it will create another EV server in Environment.
  • Hello.

    I think we misunderstood each other last time, because I installed absolutely new infrastructure at my home lab, and there is also not possible to find items against content in ZIP.

    May I explain my test case, and you will try to re-produced it in your environments. If after this test case you tell "it works", I will close this thread.

     

    1. Create *.docx file with one unique word: "unique_283_qwerty"

    2. Put this file in ZIP with native Windows zip solution: File1.zip

    3. Create *.txt file with one unique word : "asdf_496_BB"

    4. Put this file in ZIP with native Windows zip solution: File2.zip.

    5. Send mail with attachment File1.zip to recipient.

    6. Send mail with attachment File2.zip to recipient.

    7. Be sure that these messages are located in journal mailbox.

    8. Be sure that your journal mailbox is empty.

    9. Launch DA\create case\create search. Target - Vault Store which contain only journal archive.

    9.1. Content: unique_283_qwerty

    9.2. Do you receive some results?

    9.3. Content: asdf_496_BB

    9.4. Do you receive some results?

     

    Thanks.

  • Greetings, zubkoff.s;

    First, I have to ask if you are enclosing unique_283_qwerty and asdf_496_BB in double quotes (not smart quotes that you get from MS Word but normal straigh quotes)?  If not, you won't get the results you want anyway as EV indexing takes characters like the underscores and considers them to be punctuation which is not indexed.  Without the double quotes, your search criteria would be something like "unique OR 283 OR qwerty" and "asdf OR 496 OR BB".

    Also, is your indexing level on the archive being searched set to high?  If not, you won't get phrase searching either.  By default, journal archives are supposed to be set to the high indexing level, but some customers have been known to change the indexing level to low to save on index location disk space usage.

    According to the information above that you provided to Ajay, you're running EV 9.0 SP2.  Have you applied the hotfix for the EVConverterSandbox utility that may correct this issue for you? If not, please review and download the hotfix in TECH182074, "Hotfix for Symanted Enterprise Vault (EV) 9.0.2 - Converter Updates.", available at http://www.symantec.com/docs/TECH182074.  I recommend putting this into your lab and testing first.  If it resolved the issue in your lab, install it in your production environment after obtaining the appropriate Change Management approvals.  Note that this hotfix will not be retro-active to items already archived.  It will only work with new items unless you recall and re-archive to a new archive the existing items.



     

     

  • Dear Keneth Adams.

     

    1. Probably it wasn't good idea to use this unique words. But you may use more simple words, for example example123, without any special characters, underscores and quotes. But your information will be very helpful for me in a future.

     

    2. Sure, I've checked this configuration several times: index level on all journal archives set as Full.

     

    3. This update looks good, but now I am not sure that it will help me, because in my home lab I have EV 10, which has to include all latest updates for EV9. But I am going to install this update in test environment at work.

     

    Anyway, you haven't doubts that this functionality really works. It is probably misconfiguration in my environment.

    Did you try to re-produce my cases with more simple unique words?

     

    Thanks.

  • Unfortunately, I've not had time to reproduce the issue myself.  I'll work on the repro using your steps as soon as I can make the time, then report back with my results.  My testing will be in EV and DA 10.0 SP4.

  • Hi,

    Do you have any updates on this thread? Do you need more assistance regarding this topic?

    If not then please mark the post that best solves your problem as the answer to this thread.

  • Production team has confirmed the same system behavior in production environment: they didn't find the messages in DA, based on some words in ZIP files. 

    So, my test infrastructure, home lab and production environment have absolutely different configuration settings, version, hot fixes, patches, BUT have the same result with searches in ZIP files. 

    It doesn't look like coincidence. At least I haven't got clear opposite evidence from community.  

    Thanks for comment.