You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for fastar! We have a backupserver, and loads of small files, and after upgrading the hardware, I got really disappointed by 'tar' when the speed didn't increase as much as I was expecting.
It hit me that we didn't utilize the 'queue depth' of the disks when running tar. And then I found fastar, from a post on StackExchange.
But, I now suspect that 'fastar' don't respect hardlinks. I'm just running a test, and so far I see that the test have copied a lot more data than the original data I'm copying. And our data consists of insane amount of hardlinks (don't ask why.... ;) ... We're doing a kind of virtualization, and utilizing hardlinks of all files which are the same across virtual machines).
Would it be possible to get 'fastar' to understand hardlinks and keep them? Or would that not be possible with the way 'fastar' is made?
The text was updated successfully, but these errors were encountered:
Hi.
Thanks for fastar! We have a backupserver, and loads of small files, and after upgrading the hardware, I got really disappointed by 'tar' when the speed didn't increase as much as I was expecting.
It hit me that we didn't utilize the 'queue depth' of the disks when running tar. And then I found fastar, from a post on StackExchange.
But, I now suspect that 'fastar' don't respect hardlinks. I'm just running a test, and so far I see that the test have copied a lot more data than the original data I'm copying. And our data consists of insane amount of hardlinks (don't ask why.... ;) ... We're doing a kind of virtualization, and utilizing hardlinks of all files which are the same across virtual machines).
Would it be possible to get 'fastar' to understand hardlinks and keep them? Or would that not be possible with the way 'fastar' is made?
The text was updated successfully, but these errors were encountered: