[opensuse] Re: btrfs qgroup / quotas concerns - revisited
On 01/03/2017 11:11 AM, Aleksa Sarai wrote:
I am sending this message because I think it is necessary to rethink this decision. A very annoying bug that I was having in all my openSUSE machines (Leap and Tumbleweed) is actually caused by quotas. Every week, when the maintenance script is started, my systems become unresponsive for almost 30 min. I mentioned it here but received just one answer:
https://lists.opensuse.org/opensuse-factory/2016-09/msg00130.html
This also happens to me on TW, and is incredibly frustrating. So can we please also disable quotas on TW as well?
Or we could give the kernel maintainers ... kernel maintainers != brtfs developers; brtfs isn't in maintenance mode and likely won't be for some time. The problem seems caused by a user-mode script running. If that can lock up the kernel, wouldn't
Simon Lees wrote: that be good for a denial-of-service attack? But why give them any time if it is a bad decision? Why should quotas be on at all on a user system? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 3 January 2017 at 06:22, L A Walsh <suse@tlinx.org> wrote:
kernel maintainers != brtfs developers;
openSUSE's btrfs developers are also some of openSUSE's kernel maintainers. There is no point splitting hairs when you're talking about same actual people.
But why give them any time if it is a bad decision? Why should quotas be on at all on a user system?
They are enabled in order to facilitate the "Cleanup based on Disk Usage" feature in snapper - http://snapper.io/2016/05/18/space-aware-cleanup.html Considering "snapper made too many snapshots and filled up my hard disk" is by far the #1 complaint from openSUSE users from the last few years, I would say that the features required to mitigate or remove this problem are most certainly required to be on by default on a user system. Regards, Richard -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
They are enabled in order to facilitate the "Cleanup based on Disk Usage" feature in snapper - http://snapper.io/2016/05/18/space-aware-cleanup.html
Considering "snapper made too many snapshots and filled up my hard disk" is by far the #1 complaint from openSUSE users from the last few years, I would say that the features required to mitigate or remove this problem are most certainly required to be on by default on a user system.
and that would be good but TW default SPACE_LIMIT is set at 50%! over-the-top suse defaults have created and urban legend around the excesses of snapper. and all to achieve what in the typical use case? this is funny [http://www.linux-magazine.com/Issues/2015/176/Snapper] "If you leave the default settings in the /etc/snapper/configs/root file, the snapshots will quickly consume a huge amount of space. A root partition, which can normally make do with 30GB, will need between 100 and 300GB of disk space for snapshots" -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 3 January 2017 at 10:45, nicholas <ndcunliffe@gmail.com> wrote:
They are enabled in order to facilitate the "Cleanup based on Disk Usage" feature in snapper - http://snapper.io/2016/05/18/space-aware-cleanup.html
Considering "snapper made too many snapshots and filled up my hard disk" is by far the #1 complaint from openSUSE users from the last few years, I would say that the features required to mitigate or remove this problem are most certainly required to be on by default on a user system.
and that would be good but TW default SPACE_LIMIT is set at 50%!
What's wrong with that? Available Free space isn't doing anything. Using 50% of the available free space seems like a very sensible default. As space of the active volume increases, that available free space will decrease, leading to less snapshots being kept around
over-the-top suse defaults have created and urban legend around the excesses of snapper. and all to achieve what in the typical use case?
You really need to ask that question? Snapper ensures the preservation of system state before and after any user makes any change in YaST or zypper in order to be able to rollback or selectively revert anything untoward that happened either during or after a user did something in YaST or zypper. IOW snapper's saves users from incompatible packages, packaging mistakes, software changing in ways the user doesn't like, user mistakes like choosing poor settings or doing stupid stuff like "rm -Rf" in the wrong folder An openSUSE system running snapper that works today is almost guaranteed that it can also work tomorrow, because of the certainty that you can rollback to todays state in the event of changes that break it.
this is funny [http://www.linux-magazine.com/Issues/2015/176/Snapper] "If you leave the default settings in the /etc/snapper/configs/root file, the snapshots will quickly consume a huge amount of space. A root partition, which can normally make do with 30GB, will need between 100 and 300GB of disk space for snapshots"
Not bad advice, but with the new snapshot cleanup by space feature a smaller root partition is again viable as we have a way of cleaning up based on available space, or snapper only keeps a number of snapshots which it has room for, not as many as it pleases. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 03/01/2017 à 11:04, Richard Brown a écrit :
Not bad advice, but with the new snapshot cleanup by space feature a smaller root partition is again viable as we have a way of cleaning up based on available space, or snapper only keeps a number of snapshots which it has room for, not as many as it pleases.
very nice to know. May be this feature should have been emphasized a bit more. I was pretty sure something was done to make things better. Thanks jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Richard Brown wrote:
On 3 January 2017 at 10:45, nicholas <ndcunliffe@gmail.com> wrote:
They are enabled in order to facilitate the "Cleanup based on Disk Usage" feature in snapper - http://snapper.io/2016/05/18/space-aware-cleanup.html
Considering "snapper made too many snapshots and filled up my hard disk" is by far the #1 complaint from openSUSE users from the last few years, I would say that the features required to mitigate or remove this problem are most certainly required to be on by default on a user system.
and that would be good but TW default SPACE_LIMIT is set at 50%!
What's wrong with that? Available Free space isn't doing anything. Using 50% of the available free space seems like a very sensible default.
---- Not entirely. If free space < 25%, it should stop keeping snapshots -- OR, it should allocate some fixed percentage of overall disk space when it is initialized -- like MS's ability to tune the amount of space for snapshots -- defaulting to ~10% or so. If you have it dynamically using 50% of the free space, how will it respond if a user goes from 25% usage to 75% usage on their disk in one day (they created a several VM's maybe). If the space used by the snaps is contig, then not so much of a problem, but if it constantly shrinks and extends leaving fragments in the remaining free space, it will eventually hurt performance on large writes as well as overall read speed as free space becomes more and more fragmented over time.
over-the-top suse defaults have created and urban legend around the excesses of snapper. and all to achieve what in the typical use case?
An openSUSE system running snapper that works today is almost guaranteed that it can also work tomorrow, because of the certainty that you can rollback to todays state in the event of changes that break it.
---- Such is not true on MS's snapshot system. How can it be guaranteed on suse's? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 3 January 2017 at 19:58, L A Walsh <suse@tlinx.org> wrote:
Not entirely. If free space < 25%, it should stop keeping snapshots -- OR, it should allocate some fixed percentage of overall disk space when it is initialized -- like MS's ability to tune the amount of space for snapshots -- defaulting to ~10% or so.
If you have it dynamically using 50% of the free space, how will it respond if a user goes from 25% usage to 75% usage on their disk in one day (they created a several VM's maybe). If the space used by the snaps is contig, then not so much of a problem, but if it constantly shrinks and extends leaving fragments in the remaining free space, it will eventually hurt performance on large writes as well as overall read speed as free space becomes more and more fragmented over time.
I thought the default was 50% of the total space (excluding unusable space) on the root partition? That's what I read the Leap documentation on Snapper[1] as saying, as well as the documentation for the SPACE_LIMIT variable[2]. [1]: https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.ref... [2]: http://snapper.io/manpages/snapper-configs.html - Karl Cheng (Qantas94Heavy) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Karl Cheng wrote:
I thought the default was 50% of the total space (excluding unusable space) on the root partition?
!!! Yikes!... I just checked what I have set for snaps on Win7. It's currently set to 3% of the disk size or 25.84GB. That's twice the size of my linux root. That's way insane depending on your root partition size. If it used 50% of the space on my Winbox, that's be close to 500GB. If it isn't pruned continuously a bit at a time, that would explain why people are complaining about performance problems. Most people can't move around 500GB of data on their system without experiencing a noticeable performance degradation. It really should be aimed at some percentage with a maximum of 10-20G. Is it really set for 50% -- and suse has, at times advised to put everything on 1 partition, if people are using a 500-1000GB disk for suse, that could seriously impact interactive use, on say 1 hard disk. Even an SSD would be a bit stressed at moving around that amount of data. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
I thought the default was 50% of the total space (excluding unusable space) on the root partition? That's what I read the Leap documentation on Snapper[1] as saying, as well as the documentation for the SPACE_LIMIT variable[2].
[1]: https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.re ference/cha.snapper.html#sec.snapper.clean-up [2]: http://snapper.io/manpages/snapper-configs.html
- Karl Cheng (Qantas94Heavy)
That was my interpretaton: /etc/snapper/configs/root comment seems to nail it: "# fraction of the filesystems space the snapshots may use" -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (5)
-
jdd -
Karl Cheng -
L A Walsh -
nicholas -
Richard Brown