Re: [linux-lvm] Cpy%Sync column - high disk activity after reboots

Re: [linux-lvm] Cpy%Sync column - high disk activity after reboots

[Mathijs Kwik]

I'm sorry if I was unclear. But I'm not using writeback mode yet.
The period of disk activity after reboots occurs with writethrough mode.


> Hi
> 
> Cache is always flushing it's content to disk - so if you have the
> 'writeback' variant and there were some unwritten data which sits only
> in cache, they are automatically pushed to origin (cached) device -
> which is likely your disk activity.
> 
> Zdenek

Re: [linux-lvm] Cpy%Sync column - high disk activity after reboots

[Zdenek Kabelac]

Dne 2.6.2015 v 09:35 Mathijs Kwik napsal(a):
> I'm sorry if I was unclear. But I'm not using writeback mode yet.
> The period of disk activity after reboots occurs with writethrough mode.
>
>

Please attach:

dmsetup  table

dmsetup status

dmsetup ls --tree


Zdenek

[linux-lvm] Cpy%Sync column - high disk activity after reboots

[Mathijs Kwik]

Hi all,

I'm trying lvmcache and it seems to work ok. However, after every
reboot, I notice quite a long period of disk activity. lvs gives me this
output:

% lvs
  LV            VG       Attr       LSize   Pool    Origin                Data%  Meta%  Move Log Cpy%Sync Convert
  btrfs-spinner bluebook Cwi-aoC--- 582,19g [cache] [btrfs-spinner_corig] 70,11  11,34           82,55           

The % gets lower over time and when it reaches 0 the disk activity goes
back to normal.

Is this to be expected/normal?
It seems to me the entire cache is checked as if it wasn't cleanly
closed on shutdown. I'm currently using writethrough and would like to
make sure everything is ok before enabling writeback.

Thanks,
Mathijs

Re: [linux-lvm] Cpy%Sync column - high disk activity after reboots

[Zdenek Kabelac]

Dne 31.5.2015 v 16:51 Mathijs Kwik napsal(a):
> Hi all,
>
> I'm trying lvmcache and it seems to work ok. However, after every
> reboot, I notice quite a long period of disk activity. lvs gives me this
> output:
>
> % lvs
>    LV            VG       Attr       LSize   Pool    Origin                Data%  Meta%  Move Log Cpy%Sync Convert
>    btrfs-spinner bluebook Cwi-aoC--- 582,19g [cache] [btrfs-spinner_corig] 70,11  11,34           82,55
>
> The % gets lower over time and when it reaches 0 the disk activity goes
> back to normal.
>
> Is this to be expected/normal?
> It seems to me the entire cache is checked as if it wasn't cleanly
> closed on shutdown. I'm currently using writethrough and would like to
> make sure everything is ok before enabling writeback.


Hi

Cache is always flushing it's content to disk - so if you have the 'writeback' 
variant and there were some unwritten data which sits only in cache, they are 
automatically pushed to origin (cached) device - which is likely your disk 
activity.

Zdenek