Commit Graph

36 Commits

Author SHA1 Message Date
Dominik Csapak
436a48d611 restore-daemon: start disk initialization in parallel to the api
this way, the vm can start up faster, and the actual disk init happens
in parallel. this avoids unnecessary timeouts when starting the vm

if the call panics, we still abort the vm with an error

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2022-04-27 19:18:44 +02:00
Wolfgang Bumiller
99f09fd3c1 bump proxmox-compression dependency to 0.1.1
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-04-13 09:37:20 +02:00
Wolfgang Bumiller
ebb85c1ca3 bump proxmox-schema dependency to 1.3.1 for streaming attribute
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-04-13 08:20:27 +02:00
Wolfgang Bumiller
fb6e48f402 bump proxmox-router dependency to 1.2
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-04-13 08:17:08 +02:00
Thomas Lamprecht
429bc9d0a2 restore daemon: rust fmt
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2022-04-10 17:47:20 +02:00
Wolfgang Bumiller
bd4562e4b1 bump proxmox-schema dep to 1.3
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-03-04 09:50:21 +01:00
Wolfgang Bumiller
9c890d72b9 bump proxmox-async dep to 0.4
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-02-21 14:25:37 +01:00
Dominik Csapak
b066586a47 depend on new 'proxmox-compression' crate
the compression utilities live there now

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2022-02-21 14:23:43 +01:00
Fabian Grünbichler
dcf5a0f62d misc clippy fixes
the trivial ones ;)

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2022-02-08 14:57:16 +01:00
Dietmar Maurer
d91a0f9fc9 Set MMAP_THRESHOLD to a fixed value (128K)
glibc's malloc has a misguided heuristic to detect transient allocations that
will just result in allocation sizes below 32 MiB never using mmap.

That it turn means that those relatively big allocations are on the heap where
cleanup and returning memory to the OS is harder to do and easier to be blocked
by long living, small allocations at the top (end) of the heap.

Observing the malloc size distribution in a file-level backup run:

@size:
[0]                   14 |                                                    |
[1]                25214 |@@@@@                                               |
[2, 4)              9090 |@                                                   |
[4, 8)             12987 |@@                                                  |
[8, 16)            93453 |@@@@@@@@@@@@@@@@@@@@                                |
[16, 32)           30255 |@@@@@@                                              |
[32, 64)          237445 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@|
[64, 128)          32692 |@@@@@@@                                             |
[128, 256)         22296 |@@@@                                                |
[256, 512)         16177 |@@@                                                 |
[512, 1K)           5139 |@                                                   |
[1K, 2K)            3352 |                                                    |
[2K, 4K)             214 |                                                    |
[4K, 8K)            1568 |                                                    |
[8K, 16K)             95 |                                                    |
[16K, 32K)          3457 |                                                    |
[32K, 64K)          3175 |                                                    |
[64K, 128K)          161 |                                                    |
[128K, 256K)         453 |                                                    |
[256K, 512K)          93 |                                                    |
[512K, 1M)            74 |                                                    |
[1M, 2M)             774 |                                                    |
[2M, 4M)             319 |                                                    |
[4M, 8M)             700 |                                                    |
[8M, 16M)             93 |                                                    |
[16M, 32M)            18 |                                                    |

We see that all allocations will be on the heap, and that while most
allocations are small, the relatively few big ones will still make up most of
the RSS and if blocked from being released back to the OS result in much higher
peak and average usage for the program than actually required.

Avoiding the "dynamic" mmap-threshold increasement algorithm and fixing it at
the original default of 128 KiB reduces RSS size by factor 10-20 when running
backups. As with memory mappings other mappings or the heap can never block
freeing the memory fully back to the OS.

But, the drawback of using mmap is more wasted space for unaligned or small
allocation sizes, and the fact that the kernel allegedly zeros out the data
before giving it to user space. The former doesn't really matter for us when
using it only for allocations bigger than 128 KiB, and the latter is a
trade-off, using 10 to 20 times less memory brings its own performance
improvement possibilities for the whole system after all ;-)

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
 [ Thomas: added to comment & commit message + extra-empty-line fixes ]
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2022-01-26 14:10:54 +01:00
Fabian Grünbichler
6f1c26b083 tree-wide: is_ok/is_err()
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-30 15:02:07 +01:00
Fabian Grünbichler
3afecb8409 tree-wide: use is_empty() and similar
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-30 15:02:07 +01:00
Fabian Grünbichler
540fca5c9e tree-wide: cleanup manual map/flatten
found with clippy, best viewed with `-w` ;)

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-30 15:02:07 +01:00
Fabian Grünbichler
aa174e8e8a tree-wide: drop redundant clones
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-30 15:02:07 +01:00
Fabian Grünbichler
9a37bd6c84 tree-wide: fix needless borrows
found and fixed via clippy

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-30 13:55:33 +01:00
Wolfgang Bumiller
c3b8e74fdf bump regex dep to 1.5
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2021-12-16 11:25:02 +01:00
Wolfgang Bumiller
821aa8eae6 bump proxmox-schema to 1.1
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2021-12-16 11:25:02 +01:00
Thomas Lamprecht
25e41aa802 restore-daemon: fix use of deprecated env_logger::from_env function
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2021-12-16 11:12:36 +01:00
Thomas Lamprecht
ff6b6cd74d drop unused imports of proxmox_sys::identity
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2021-12-16 11:09:38 +01:00
Fabian Grünbichler
b25e07b3ce deps: env_logger update to 0.9
and removal from main crate, not needed there anymore.

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-12-01 12:46:37 +01:00
Dietmar Maurer
25877d05ac update to proxmox-sys 0.2 crate
- imported pbs-api-types/src/common_regex.rs from old proxmox crate
- use hex crate to generate/parse hex digest
- remove all reference to proxmox crate (use proxmox-sys and
  proxmox-serde instead)

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
2021-11-24 10:32:27 +01:00
Dietmar Maurer
92a8f0bc82 depend on proxmox-async 0.2 2021-11-20 17:14:02 +01:00
Dietmar Maurer
9a1b24b6b1 use new proxmox-async crate
Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
2021-11-19 18:03:22 +01:00
Fabian Grünbichler
daaeea8b4b update to base64 0.13
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
2021-11-19 11:58:17 +01:00
Dietmar Maurer
860eaec58f use proxmox::tools::fd::fd_change_cloexec from proxmox 0.15.3
Depend on proxmox 0.15.3

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
2021-11-18 13:43:41 +01:00
Dietmar Maurer
e0a19d3313 use new fsync parameter to replace_file and atomic_open_or_create
Depend on proxmox 0.15.0 and proxmox-openid 0.8.1

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
2021-10-21 07:28:32 +02:00
Dietmar Maurer
b3f279e2d9 use complete_file_name from proxmox-router 1.1 2021-10-13 14:10:02 +02:00
Wolfgang Bumiller
6ef1b649d9 update to first proxmox crate split
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2021-10-11 11:58:49 +02:00
Wolfgang Bumiller
e3f3359c86 bump proxmox dependency to 0.14.0 and proxmox-http to 0.5.0
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
2021-10-08 11:18:22 +02:00
Dietmar Maurer
608806e884 proxmox-rest-server: use new ServerAdapter trait instead of callbacks
Async callbacks are a PITA, so we now pass a single trait object which
implements check_auth and get_index.
2021-10-05 11:13:10 +02:00
Dietmar Maurer
48176b0a77 proxmox-rest-server: pass owned RestEnvironment to get_index
This way we avoid pointers with lifetimes.
2021-10-05 11:12:53 +02:00
Dietmar Maurer
6680878b5c proxmox-rest-server: make get_index async 2021-10-01 09:38:10 +02:00
Dietmar Maurer
038f385089 proxmox-rest-server: make check_auth async 2021-10-01 07:53:59 +02:00
Dietmar Maurer
99940358e3 ExtJsFormatter: use ParameterError to correctly compute 'errors'
By default, 'errors' is now empty.

Depend on proxmox 0.13.5.
2021-09-28 10:19:55 +02:00
Dietmar Maurer
81867f0539 use UPID and systemd helpers from proxmox 0.13.4 2021-09-23 12:01:43 +02:00
Dietmar Maurer
6523588c8d move proxmox_restore_daemon code into extra crate
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2021-09-21 08:46:41 +02:00