From: Christian Ebner <c.ebner@proxmox.com>
To: "Proxmox Backup Server development discussion"
<pbs-devel@lists.proxmox.com>,
"Fabian Grünbichler" <f.gruenbichler@proxmox.com>
Subject: Re: [pbs-devel] [PATCH proxmox 2/2] pbs api types: add option to set GC chunk cleanup wait period
Date: Mon, 3 Mar 2025 14:30:00 +0100 [thread overview]
Message-ID: <f57a9703-072c-49bf-b1be-6a761775c4ac@proxmox.com> (raw)
In-Reply-To: <1741005633.eo593uhjhw.astroid@yuna.none>
On 3/3/25 13:58, Fabian Grünbichler wrote:
> On February 19, 2025 5:48 pm, Christian Ebner wrote:
>> Add the `gc-wait-period` option to the datastore tuning parameters.
>> This allows to specify the time after which the chunks are not
>> considered in use anymore if their atime has not been updated since
>> then. This option is only considered, if the `gc-atime-check` is
>> enabled, to avoid potential data loss.
>>
>> The default is to keep chunks within the 24h 5m timespan (given no
>> active writers). The 5m minimum was chosen to be in line with the
>> already used safety offset for garbage collection.
>>
>> Signed-off-by: Christian Ebner <c.ebner@proxmox.com>
>> ---
>> changes since version 1:
>> - not present in previous version
>>
>> pbs-api-types/src/datastore.rs | 13 +++++++++++++
>> 1 file changed, 13 insertions(+)
>>
>> diff --git a/pbs-api-types/src/datastore.rs b/pbs-api-types/src/datastore.rs
>> index a7a9db7d..7542e969 100644
>> --- a/pbs-api-types/src/datastore.rs
>> +++ b/pbs-api-types/src/datastore.rs
>> @@ -223,6 +223,13 @@ pub enum DatastoreFSyncLevel {
>> Filesystem,
>> }
>>
>> +pub const GC_WAIT_PERIOD_SCHEMA: Schema =
>> + IntegerSchema::new("Wait period (in minutes) for garbage collection phase 2 chunk cleanup (default 24h 5m)")
>> + .minimum(5)
>> + .maximum(1445)
>
> these seem a bit conservative - if we introduce the option, we could
> also allow to reduce the grace period to 0 or to require longer grace
> periods?
I did stick to the additional 5 minutes safety gap, as this was also
introduced with the oldest writer change in commit 11861a48
("src/backup/chunk_store.rs: fix GC"). Since that commit does not
explicitly mention the reason (I assume to prevent a race between writer
instantiation and garbage collection), I sticked with this additional
lower limit.
Increasing the upper limit might be possible without such concerns, I do
not really see however the use case for that.
>
> wait period is also a bit of a misnomer IMHO, this is not something
> where we "wait" per se, but rather something to account for the
> timestamp information potentially not being accurate.. not sure what a
> better term would be, "grace period" doesn't really fit either..
I sticked with Thomas' naming suggestion here, but maybe something like
`gc-clenup-safety-margin` is more fitting? Although it does not imply
that this is due to the atime... Other suggestions?
>
> it's a kind of "safety margin", maybe something related to that? naming
> can be quite hard..
>
> maybe gc_atime_safety_margin ? that also implies that setting a
> non-default value has potential safety implications/risks your data,
> IMHO more than "reducing a wait period" does..
>
>> + .default(1445)
>> + .schema();
>> +
>> #[api(
>> properties: {
>> "chunk-order": {
>> @@ -235,6 +242,10 @@ pub enum DatastoreFSyncLevel {
>> default: true,
>> type: bool,
>> },
>> + "gc-wait-period": {
>> + schema: GC_WAIT_PERIOD_SCHEMA,
>> + optional: true,
>> + },
>> },
>> )]
>> #[derive(Serialize, Deserialize, Default)]
>> @@ -248,6 +259,8 @@ pub struct DatastoreTuning {
>> pub sync_level: Option<DatastoreFSyncLevel>,
>> #[serde(skip_serializing_if = "Option::is_none")]
>> pub gc_atime_check: Option<bool>,
>> + #[serde(skip_serializing_if = "Option::is_none")]
>> + pub gc_wait_period: Option<usize>,
>> }
>>
>> pub const DATASTORE_TUNING_STRING_SCHEMA: Schema = StringSchema::new("Datastore tuning options")
>> --
>> 2.39.5
>
>
> _______________________________________________
> pbs-devel mailing list
> pbs-devel@lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
>
>
_______________________________________________
pbs-devel mailing list
pbs-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
next prev parent reply other threads:[~2025-03-03 13:30 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-19 16:48 [pbs-devel] [PATCH proxmox proxmox-backup 0/9] fix #5892: check atime update is honored Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox 1/2] pbs api types: add check garbage collection atime updates flag Christian Ebner
2025-03-03 12:58 ` Fabian Grünbichler
2025-03-03 13:20 ` Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox 2/2] pbs api types: add option to set GC chunk cleanup wait period Christian Ebner
2025-03-03 12:58 ` Fabian Grünbichler
2025-03-03 13:30 ` Christian Ebner [this message]
2025-03-03 15:30 ` Fabian Grünbichler
2025-03-04 16:01 ` Thomas Lamprecht
2025-03-04 16:37 ` Christian Ebner
2025-03-04 16:49 ` Thomas Lamprecht
2025-03-04 17:03 ` Christian Ebner
2025-03-05 7:57 ` Fabian Grünbichler
2025-03-05 8:12 ` Fabian Grünbichler
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 3/9] datastore: move utimensat() related constants to module scope Christian Ebner
2025-03-03 13:03 ` Fabian Grünbichler
2025-03-03 13:32 ` Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 4/9] fix #5982: garbage collection: check atime updates are honored Christian Ebner
2025-03-03 13:51 ` Fabian Grünbichler
2025-03-03 15:23 ` Christian Ebner
2025-03-03 15:50 ` Fabian Grünbichler
2025-03-03 16:09 ` Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 5/9] ui: expose atime update check flag in datastore tuning options Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 6/9] docs: mention GC atime update check for " Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 7/9] datastore: conditionally use custom GC wait period if set Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 8/9] ui: expose GC wait period in datastore tuning option Christian Ebner
2025-02-19 16:48 ` [pbs-devel] [PATCH proxmox-backup 9/9] docs: mention gc-wait-period option as datastore tuning parameter Christian Ebner
2025-03-04 18:38 ` [pbs-devel] [PATCH proxmox proxmox-backup 0/9] fix #5892: check atime update is honored Christian Ebner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=f57a9703-072c-49bf-b1be-6a761775c4ac@proxmox.com \
--to=c.ebner@proxmox.com \
--cc=f.gruenbichler@proxmox.com \
--cc=pbs-devel@lists.proxmox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.