From: Dietmar Maurer <dietmar@proxmox.com>
To: Proxmox Backup Server development discussion
<pbs-devel@lists.proxmox.com>,
Dominik Csapak <d.csapak@proxmox.com>
Subject: Re: [pbs-devel] [RFC PATCH proxmox-backup 1/2] tape: media_catalog: commit as much as possible
Date: Wed, 21 Jul 2021 13:18:31 +0200 (CEST) [thread overview]
Message-ID: <482430901.322.1626866312005@webmail.proxmox.com> (raw)
this looks clumsy.
Instead, can we make catalog.start_chunk_archive and catalog.end_chunk_archive() private,
and add a new method to register whole archives (catalog.register_chunk_archive())
That way all pending changes are valid and can be commited?
> On 07/20/2021 11:56 AM Dominik Csapak <d.csapak@proxmox.com> wrote:
>
>
> record when we finish an archive in the catalogs pending data,
> and commit data until there. This way we can commit the catalog
> even if we do not finish a chunk archive.
>
> Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
> ---
> sending as RFC, since i am not completely sure if i thought this
> through completely, but in my tests it worked..
>
> src/tape/media_catalog.rs | 31 ++++++++++++++++++++++---------
> 1 file changed, 22 insertions(+), 9 deletions(-)
>
> diff --git a/src/tape/media_catalog.rs b/src/tape/media_catalog.rs
> index 65b52a42..8dacf986 100644
> --- a/src/tape/media_catalog.rs
> +++ b/src/tape/media_catalog.rs
> @@ -66,6 +66,7 @@ pub struct MediaCatalog {
>
> content: HashMap<String, DatastoreContent>,
>
> + finished_archive_len: usize,
> pending: Vec<u8>,
> }
>
> @@ -223,6 +224,7 @@ impl MediaCatalog {
> current_archive: None,
> last_entry: None,
> content: HashMap::new(),
> + finished_archive_len: 0,
> pending: Vec::new(),
> };
>
> @@ -231,6 +233,7 @@ impl MediaCatalog {
> if !found_magic_number {
> me.pending.extend(&Self::PROXMOX_BACKUP_MEDIA_CATALOG_MAGIC_1_1);
> }
> + me.finished_archive_len = me.pending.len();
>
> if write {
> me.file = Some(file);
> @@ -300,12 +303,15 @@ impl MediaCatalog {
> current_archive: None,
> last_entry: None,
> content: HashMap::new(),
> + finished_archive_len: 0,
> pending: Vec::new(),
> };
>
> me.log_to_stdout = log_to_stdout;
>
> me.pending.extend(&Self::PROXMOX_BACKUP_MEDIA_CATALOG_MAGIC_1_1);
> + me.finished_archive_len = me.pending.len();
> +
>
> me.register_label(&media_id.label.uuid, 0, 0)?;
>
> @@ -366,31 +372,34 @@ impl MediaCatalog {
> ///
> /// Fixme: this should be atomic ...
> pub fn commit(&mut self) -> Result<(), Error> {
> -
> - if self.pending.is_empty() {
> + if self.finished_archive_len == 0 {
> return Ok(());
> }
>
> match self.file {
> Some(ref mut file) => {
> - file.write_all(&self.pending)?;
> + file.write_all(&self.pending[0..self.finished_archive_len])?;
> file.flush()?;
> file.sync_data()?;
> }
> None => bail!("media catalog not writable (opened read only)"),
> }
>
> - self.pending = Vec::new();
> + let pending = self.pending.len();
> + if self.finished_archive_len < pending {
> + self.pending
> + .copy_within(self.finished_archive_len..pending, 0);
> + }
> + self.pending
> + .truncate(self.pending.len() - self.finished_archive_len);
> + self.finished_archive_len = 0;
>
> Ok(())
> }
>
> - /// Conditionally commit if in pending data is large (> 1Mb)
> + /// Conditionally commit if finished archives in pending data is large (> 1MiB)
> pub fn commit_if_large(&mut self) -> Result<(), Error> {
> - if self.current_archive.is_some() {
> - bail!("can't commit catalog in the middle of an chunk archive");
> - }
> - if self.pending.len() > 1024*1024 {
> + if self.finished_archive_len > 1024*1024 {
> self.commit()?;
> }
> Ok(())
> @@ -498,6 +507,7 @@ impl MediaCatalog {
>
> unsafe { self.pending.write_le_value(entry)?; }
>
> + self.finished_archive_len = self.pending.len();
> self.last_entry = Some((uuid.clone(), file_number));
>
> Ok(())
> @@ -625,6 +635,7 @@ impl MediaCatalog {
>
> unsafe { self.pending.write_le_value(entry)?; }
>
> + self.finished_archive_len = self.pending.len();
> self.last_entry = Some((uuid, file_number));
> }
> }
> @@ -688,6 +699,8 @@ impl MediaCatalog {
> self.pending.push(b':');
> self.pending.extend(snapshot.as_bytes());
>
> + self.finished_archive_len = self.pending.len();
> +
> let content = self.content.entry(store.to_string())
> .or_insert(DatastoreContent::new());
>
> --
> 2.30.2
>
>
>
> _______________________________________________
> pbs-devel mailing list
> pbs-devel@lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
next reply other threads:[~2021-07-21 11:18 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-21 11:18 Dietmar Maurer [this message]
-- strict thread matches above, loose matches on Subject: below --
2021-07-20 9:56 Dominik Csapak
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=482430901.322.1626866312005@webmail.proxmox.com \
--to=dietmar@proxmox.com \
--cc=d.csapak@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.