Our future backup code: test it!

florent

@Tristis-Oris did you run yarn from the root of the project to install the new dependencies ?

Tristis Oris

yarn install v1.22.22
[1/5] Validating package.json...
[2/5] Resolving packages...
[3/5] Fetching packages...
warning Pattern ["@intlify/shared@latest"] is trying to unpack in the same destination "/usr/local/share/.cache/yarn/v6/npm-@intlify-shared-11.1.2-d780bc8eb4e3c3ef8a3e45c421f0f5ecf59651f2-integrity/node_modules/@intlify/shared" as pattern ["@intlify/shared@11.1.2","@intlify/shared@11.1.2","@intlify/shared@11.1.2"]. This could result in non-deterministic behavior, skipping.
warning Pattern ["@intlify/shared@next"] is trying to unpack in the same destination "/usr/local/share/.cache/yarn/v6/npm-@intlify-shared-11.0.0-rc.1-52a67aa12fccd9303027b48ebf017b75b7350283-integrity/node_modules/@intlify/shared" as pattern ["@intlify/shared@11.0.0-rc.1"]. This could result in non-deterministic behavior, skipping.
warning bare-os@3.4.0: The engine "bare" appears to be invalid.
[4/5] Linking dependencies...
warning "@commitlint/cli > @commitlint/load > cosmiconfig-typescript-loader@6.1.0" has unmet peer dependency "@types/node@*".
warning "@commitlint/cli > @commitlint/load > cosmiconfig-typescript-loader@6.1.0" has unmet peer dependency "typescript@>=5".
warning "@typescript-eslint/eslint-plugin > ts-api-utils@1.4.3" has unmet peer dependency "typescript@>=4.2.0".
warning " > @xen-orchestra/web-core@0.18.0" has unmet peer dependency "pinia@^3.0.1".
warning " > @xen-orchestra/web-core@0.18.0" has unmet peer dependency "vue@~3.5.13".
warning " > @xen-orchestra/web-core@0.18.0" has unmet peer dependency "vue-i18n@^11.1.2".
warning " > @xen-orchestra/web-core@0.18.0" has unmet peer dependency "vue-router@^4.5.0".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @xen-orchestra/rest-api > inversify@6.2.2" has unmet peer dependency "reflect-metadata@~0.2.2".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @xen-orchestra/rest-api > swagger-ui-express@5.0.1" has unmet peer dependency "express@>=4.0.0 || >=5.0.0-beta".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @vates/event-listeners-manager > tap > @tapjs/test > @isaacs/ts-node-temp-fork-for-pr-2009@10.9.7" has unmet peer dependency "@types/node@*".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @xen-orchestra/rest-api > inversify > @inversifyjs/core > @inversifyjs/reflect-metadata-utils@0.2.4" has unmet peer dependency "reflect-metadata@0.2.2".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @vates/event-listeners-manager > tap > @tapjs/asserts > tcompare > react-element-to-jsx-string@15.0.0" has unmet peer dependency "react@^0.14.8 || ^15.0.1 || ^16.0.0 || ^17.0.1 || ^18.0.0".
warning "workspace-aggregator-9e56f98d-aa3d-4c5b-aae2-45a491674f9d > @vates/event-listeners-manager > tap > @tapjs/asserts > tcompare > react-element-to-jsx-string@15.0.0" has unmet peer dependency "react-dom@^0.14.8 || ^15.0.1 || ^16.0.0 || ^17.0.1 || ^18.0.0".
warning Workspaces can only be enabled in private projects.
[5/5] Building fresh packages...
$ husky install
husky - Git hooks installed
Done in 39.86s.

but same error.

florent

@Tristis-Oris it looks like the build from completly empty repo don't work as expected

can you try this :

cd @vates/generator-toolbox/
yarn build 
cd - 
yarn build

I am working on improving the build process in the meantimes

Tristis Oris

@florent yep it works now.

john.c

So is it this new backup code which the tape back up support will be plugging into?

If so I have been thinking may be have it when doing the tape backup support, get the XOA to install the packages when required after prompting the user. Also have xo-server instruct the creation of a new disk image for attaching to xoa. The new disk image would act as storage or cache for the extra data used by tape drive backups.

The tape drive related software wouldn't be installed by default, unless specified as such during deployment with the XOA deploy tool. But instead installed at user request, by selecting to use tape backup destination when setting up back up job, clicking on the shim widget about tape backups or restoring from an online Vates account backup of XOA settings.

Davidj 0

@florent I'm doing this from source.
Running git pull --rebase seems to have fixed the build problems.

florent

@john.c the backup on tape is still on the roadmap but won't be for this step

The main change is that we'll have to rework part of the code to ensure we write data sequentially + the management tools ( which backup is on which tape )

You can try the FS on tape , but the performance penalty is heavy

john.c

@florent said in Our future backup code: test it!:

@john.c the backup on tape is still on the roadmap but won't be for this step

The main change is that we'll have to rework part of the code to ensure we write data sequentially + the management tools ( which backup is on which tape )

You can try the FS on tape , but the performance penalty is heavy

Even if its not part of this rewrite some of what I wrote is my idea or view as to how it could be implemented later on. So that it's implemented in a way, that permits people who don't need tape back ups to still have small Xen Orchestra instances. But at the same time enable those who use tape backups, to have the capacity and an appropriately sized Xen Orchestra.

Davidj 0

Testing a continuous replication backup, I get Error: can't connect to any nbd client

Running the same job on my regular XO works fine. Let me know what logs you want to see.

mail (h2)

   Snapshot
   Start: 2025-04-10 18:18
   End: 2025-04-10 18:18
   h3 local storage (130.02 GiB free) - h3
   Start: 2025-04-10 18:18

Start: 2025-04-10 18:18
End: 2025-04-10 18:18
Duration: a few seconds
Error: can't connect to any nbd client
Type: delta

Tristis Oris

@Tristis-Oris probably after https://github.com/vatesfr/xen-orchestra/pull/8432/files/04be6032180e7245a8db9c5f78177ac7999882c2

/@vates/generator-toolbox# yarn build
yarn run v1.22.22
$ tsc
/bin/sh: 1: tsc: not found
error Command failed with exit code 127.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

ok, npm install -g typescript

yarn build
yarn run v1.22.22
$ tsc
src/synchronized.mts:1:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

1 import assert from 'node:assert'
                     ~~~~~~~~~~~~~

src/synchronized.test.mts:1:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

1 import assert from 'node:assert'
                     ~~~~~~~~~~~~~

src/synchronized.test.mts:2:29 - error TS2307: Cannot find module 'node:test' or its corresponding type declarations.

2 import { suite, test } from 'node:test'
                              ~~~~~~~~~~~

src/throttle.mts:1:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

1 import assert from 'node:assert'
                     ~~~~~~~~~~~~~

src/throttle.test.mts:2:29 - error TS2307: Cannot find module 'node:test' or its corresponding type declarations.

2 import { suite, test } from 'node:test'
                              ~~~~~~~~~~~

src/throttle.test.mts:3:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

3 import assert from 'node:assert'
                     ~~~~~~~~~~~~~

src/throttle.test.mts:7:11 - error TS2580: Cannot find name 'Buffer'. Do you need to install type definitions for node? Try `npm i --save-dev @types/node`.

7     yield Buffer.allocUnsafe(chunkSize)
            ~~~~~~

src/timeout.mts:1:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

1 import assert from 'node:assert'
                     ~~~~~~~~~~~~~

src/timeout.test.mts:1:30 - error TS2307: Cannot find module 'node:test' or its corresponding type declarations.

1 import { describe, it } from 'node:test'
                               ~~~~~~~~~~~

src/timeout.test.mts:2:20 - error TS2307: Cannot find module 'node:assert' or its corresponding type declarations.

2 import assert from 'node:assert'
                     ~~~~~~~~~~~~~


Found 10 errors in 6 files.

Errors  Files
     1  src/synchronized.mts:1
     2  src/synchronized.test.mts:1
     1  src/throttle.mts:1
     3  src/throttle.test.mts:2
     1  src/timeout.mts:1
     2  src/timeout.test.mts:1
error Command failed with exit code 2.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

fbeauchamp opened this pull request in vatesfr/xen-orchestra

closed feat(backups): use generator instead of streams for backup and replication #8432

Tristis Oris

@Tristis-Oris
from root directory.
/xen-orchestra
yarn - succeess
yarn build - failed
again to cd @vates/generator-toolbox/
yarn build

yarn run v1.22.22
$ tsc
Done in 5.50s.

cd -
yarn build

 Tasks:    29 successful, 29 total
Cached:    13 cached, 29 total
  Time:    1m17.02s

Done in 77.34s.

again at root

cd packages/xo-server
mkdir -p ~/.config/xo-server
cp sample.config.toml ~/.config/xo-server/config.toml
yarn start

Tristis Oris

delta, first full.

no NBD

2025-04-11T12:19:15.169Z xo:backups:worker INFO starting backup
2025-04-11T12:22:29.509Z xo:backups:MixinBackupWriter WARN cleanVm: incorrect backup size in metadata {
  path: '/xo-vm-backups/07e5d245-7ed3-b690-e386-3461556ad41f/20250411T121931Z.json',
  actual: 27271364608,
  expected: 27278188544
}

2 vms but 1 log at cli.

delta

[2025-04-11T12:27:41.428Z xo:backups:worker INFO starting backup
2025-04-11T12:27:48.644Z xo:backups:MixinBackupWriter WARN cleanVm: incorrect backup size in metadata {
  path: '/xo-vm-backups/f3cfdc19-0e6a-04d1-5651-82e1dbb25191/20250411T122746Z.json',
  actual: 44040192,
  expected: 46314496
}
2025-04-11T12:28:03.055Z xo:backups:MixinBackupWriter WARN cleanVm: incorrect backup size in metadata {
  path: '/xo-vm-backups/07e5d245-7ed3-b690-e386-3461556ad41f/20250411T122801Z.json',
  actual: 46137344,
  expected: 48412160
}
2025-04-11T12:28:03.081Z xo:backups:worker INFO backup has ended
2025-04-11T12:28:03.181Z xo:backups:worker INFO process will exit {
  duration: 21752192,
  exitCode: 0,
  resourceUsage: {
    userCPUTime: 3794057,
    systemCPUTime: 468414,
    maxRSS: 69592,
    sharedMemorySize: 0,
    unsharedDataSize: 0,
    unsharedStackSize: 0,
    minorPageFault: 42277,
    majorPageFault: 0,
    swappedOut: 0,
    fsRead: 13624,
    fsWrite: 185840,
    ipcSent: 0,
    ipcReceived: 0,
    signalsCount: 0,
    voluntaryContextSwitches: 3591,
    involuntaryContextSwitches: 441
  },
  summary: { duration: '22s', cpuUsage: '20%', memoryUsage: '67.96 MiB' }
}

can't get NBD to work. It enabled on nic, but not used for backups. XO settings copied from prod instance.

Davidj 0

Like @Tristis-Oris , my first tests were done using a clone of my production XO. They failed with "No NBD" errors.

Creating a new backup works.

Tristis Oris

@Davidj-0 No NBD usage even with new task( Main XO is on same host, same nic.

Davidj 0

@Tristis-Oris My setup is similar to yours, and I get similar errors on any existing backup job.

However, if I create a new backup job, then it works without any error.

ping @florent , maybe these data points are useful.

florent

thanks for your tests . I will recheck NBD backup

florent

ih @Tristis-Oris and @Davidj-0 I pushed an update to better handle NBD error, please keep us informed

Tristis Oris

@florent i'm not sure how to update from this branch(

git checkout .
git pull --ff-only
yarn
yarn build

Updating 04be60321..6641580e6
Fast-forward
 @xen-orchestra/xapi/disks/Xapi.mjs             | 17 ++++++++++-------
 @xen-orchestra/xapi/disks/XapiVhdCbt.mjs       |  3 ++-
 @xen-orchestra/xapi/disks/XapiVhdStreamNbd.mjs |  3 ---
 @xen-orchestra/xapi/disks/utils.mjs            | 29 ++++++++++++++++-------------
 4 files changed, 28 insertions(+), 24 deletions(-)

git checkout feat_generator_backups
Already on 'feat_generator_backups'
Your branch is up to date with 'origin/feat_generator_backups'.


git pull origin feat_generator_backups
From https://github.com/vatesfr/xen-orchestra
 * branch                feat_generator_backups -> FETCH_HEAD
Already up to date.

but still no NBD at backup.

Tristis Oris

i see significant speed difference from master branch.

prod
NBD
160.71 MiB
no NBD
169

test
NBD (broken)
110-130 MiB
no NBD
115

florent

@Tristis-Oris ouch that is quite costly
Can you describe which backup you run ?
Can you check if the duration is different (maybe this is a measurement error and not a slow speed) ?