Re: [linux-lvm] lvdisplay "open" field computing

[Simon ELBAZ <selbaz@linagora.com>]

Hi,

I have setup a Debian 10 VM to understand the issue.

root@sympa:~# uname -a
Linux sympa 4.19.0-5-amd64 #1 SMP Debian 4.19.37-5 (2019-06-19) x86_64 
GNU/Linux
root@sympa:~# lvdisplay /dev/sympa-vg/swap_1
   --- Logical volume ---
   LV Path                /dev/sympa-vg/swap_1
   LV Name                swap_1
   VG Name                sympa-vg
   LV UUID                1b59OW-M2yW-PuI1-QN5t-pN0w-6Akl-AnJnNh
   LV Write Access        read/write
   LV Creation host, time sympa, 2019-07-10 11:06:42 +0200
   LV Status              available
   # open                 2
   LV Size                <1,99 GiB
   Current LE             509
   Segments               1
   Allocation             inherit
   Read ahead sectors     auto
   - currently set to     256
   Block device           254:1

root@sympa:~# lsof | grep "254,1"
root@sympa:~# dmsetup info
Name:              sympa--vg-swap_1
State:             ACTIVE
Read Ahead:        256
Tables present:    LIVE
Open count:        2
Event number:      0
Major, minor:      254, 1
Number of targets: 1
UUID: LVM-Imuz5YW2OXfsyO2ChOPPLI5flib9YlLi1b59OWM2yWPuI1QN5tpN0w6AklAnJnNh

Name:              sympa--vg-root
State:             ACTIVE
Read Ahead:        256
Tables present:    LIVE
Open count:        1
Event number:      0
Major, minor:      254, 0
Number of targets: 1
UUID: LVM-Imuz5YW2OXfsyO2ChOPPLI5flib9YlLiJMF1vN1e6TKmfV1rT0fvWJnpViu7BMw5

root@sympa:~# lsof | grep swap_1

root@sympa:~#

Could someone pls give me some hints to verify the "open" value ?

Thanks

On 10/07/2019 15:37, Simon ELBAZ wrote:
> Thank you very much Zdenek.
>
> I will suggest the customer a kernel update.
>
> On 10/07/2019 15:32, Zdenek Kabelac wrote:
>> Dne 10. 07. 19 v 13:54 Simon ELBAZ napsal(a):
>>> Hi Zdenek,
>>>
>>> Thanks for your feedback.
>>>
>>> The kernel version is:
>>>
>>> [root@panoramix ~]# uname -a
>>> Linux panoramix.ch-perrens.fr 2.6.32-573.12.1.el6.x86_64 #1 SMP Tue 
>>> Dec 15 21:19:08 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
>>>
>>
>> Hi
>>
>> So this is really a very historical kernel - released ~10 years back.
>> I'm afraid none is going to look out for the reasoning of any kernel 
>> bug....
>>
>> It might be interesting to see if you get any reproducer to give you 
>> hints
>> how to avoid this happen (at least not easily).
>>
>>> Jul  2 03:08:10 panoramix LVM(pri_ISCSIVG0_vg_obm)[18618]: INFO: 
>>> Retry deactivating volume group vg_obm
>>>
>>> This is why I am trying to understand how the field is computed.
>>
>> Simple rule applied here: DM devices in use (open_count > 0) cannot 
>> be deactivated.
>>
>> Occasionally there were 'race events' with udev - where device
>> which should have been otherwise unused has been asynchronously
>> opened by udev scanning rules - but that's like is not your case.
>>
>> As you seems to have device open count higher permanently -
>>
>> So if you are sure there is no running 'APP' after umount, that
>> keeps device open - it could be likely be some very old bug in kernel
>> with very high probably 99.99999% of being fixed ;)
>>
>> Regards
>>
>>
>> Zdenek
>>
-- 
Simon Elbaz
@Linagora
Mob: +33 (0) 6 38 99 18 34
Tour Franklin 31ème étage
100/101 Quartier Boieldieu
92042 La Défense
FRANCE