Virtuozzo 6.0 Update 11 Hotfix 8 (6.0.11-3468) - PSBM-46563

Discussion in 'General Questions' started by futureweb, Jul 22, 2016.

  1. futureweb

    futureweb Tera Poster

    Messages:
    397
    Hey there,

    according to Release Notes of the latest VZ update the EXT4 Defrag should have gotten better. (https://kb.virtuozzo.com/en/129352 - Improved ext4 defragmentation. (PSBM-46563))

    Did update my Nodes and tested with pcompact and also with "prl_disk_tool compact" - but did not get better results than before this patch?

    Any ideas on that?

    Thx
    Andreas Schnederle-Wagner
     
  2. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    Defragmentation was improved for certain use-case. You did mention that "it didn't get better" but you have not supplied anything useful for investigation.
    It's still not clear whether what you expected to be improved is some bug, or just wrong expectation.

    Please elaborate what result did you receive and what did you expect to get?
     
  3. futureweb

    futureweb Tera Poster

    Messages:
    397
    Hey Pavel,
    as you may know - we already had a Support Ticket regarding very bad Thin Provisioning Space Reclaimation (Support Ticket #2057204). (Guess you also remember our other Thread about this: http://forum.odin.com/threads/pstorage-used-space.337587/ :))
    About 1/4 of our licensed capacity is wasted because it´s not reclaimed.

    Within Support Ticket #2057204 they refered to excactly this internal Bug Ticket (iloginovskikh - Bug was assigned to ticket: BZ link #PSBM-46563) --> emareeva - Correspondence added: This particular issue should be fixed in one of the future updates in scope of the bug PSBM-46563

    Now #PSBM-46563 was released - but pcompact & "prl_disk_tool compact" isn't reclaiming a single MB more than before this Update! :-/

    Andreas
     
  4. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    Bugfix was included into 117.8 kernel. If your server is running this kernel at the moment, and defrag behavior did not improve you should contact support for further investigation. But only if you're currently running this kernel.
     
  5. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    On a second thought, after investigating internal notes a bit further, only a part of the fix was delivered - kernel part.
    Userspace part is to be provided in one of the future hotfixes, thus it's too early to consider the bug fixed -> thus no point in submitting support ticket.
     
  6. futureweb

    futureweb Tera Poster

    Messages:
    397
    alright - thx for digging a bit deeper in the internal notes! ;-)
    Any (rough) ETA when the Userspace part will become available?
    As "lost Space" is giving me some headache! ;-)
     
  7. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    Unfortunately - no ETA so far. Work in progress
     
  8. SteveITS

    SteveITS Tera Poster

    Messages:
    271
    We received an email about Hotfix 8 containing:

    "We have just become aware of potential vulnerabilities associated with kernel version 2.6.32-042stab117.8
    We advise that you downgrade the kernel on any hosts running this version."

    The linked support site doesn't have an obvious link on how to do that. The KB site/page has been redone and doesn't sort in date order unfortunately so there's no way to find "new" articles. I did find this one which fixes a disk calculation/corruption issue for EXT4. Presumably there was a Hotfix 9 that fixed said vulnerability?

    https://help.virtuozzo.com/customer...irtuozzo-6-0-update-11-hotifx-10-6-0-11-3474-
     
  9. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    Technically speaking this kernel is the same as 117.8, but without the patch that introduced the bug. You can just install this kernel update and everything will be fine.
     
  10. futureweb

    futureweb Tera Poster

    Messages:
    397
    Hi Pavel,
    by any chance any updates / ETA on PSBM-46563 Userspace Part?
     
  11. futureweb

    futureweb Tera Poster

    Messages:
    397
    Hi Pavel,
    do you have ANY news on PSBM-46563 Userspace Part? We are loosing hundreds of GBs on our installation because of this issue ...
    thx
    Andreas
     
  12. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    There is no ETA.
    Note, vz6 has reached EOM in November.
    You can check the lifecycle and description for the stages following the link:
    https://virtuozzo.com/support/server-lifecycle/

    Although bug is unpleasant and causes CTs to use more space than they should, it's not a critical issue and does not fall into EOM scope. It's not a priority task right now.

    Moreover, "fix" itself is not trivial at all. It's not like "we can release it anytime but for some reason are hiding the patch". No, it's because the task is very difficult and requires a lot of serious research. It's not "adjust logic" task, it's "implement new rocket engine" task.
     
  13. futureweb

    futureweb Tera Poster

    Messages:
    397
    thx for clarification! ;-)
    Is this Bug still present in VZ7 or would upgrading solve this problem? (and hopefully don't bring in lot's of other problems ... hehe :) )
     
  14. Pavel

    Pavel A.I. Auto-Responder Staff Member

    Messages:
    478
    Unfortunately it's still present in vz7.
    It's difficult to call it a "bug" - more like technology flaw :) One that makes compacting not efficient enough. Addressing this flaw requires new technologies. But we'll be working on improving that in vz7.
     
  15. futureweb

    futureweb Tera Poster

    Messages:
    397
    alright - I hope your Devs develop this ""new rocket engine"" soon ;-))
     
  16. futureweb

    futureweb Tera Poster

    Messages:
    397
    Are there any News on this Topic? Now we got over 600GB of wasted space because of this ... :-/
    If this would be better in VZ7 I would try to schedule Update sooner than later ...

    thx
    Andreas
     

Share This Page