Commit Graph

1100 Commits

Author SHA1 Message Date
Christian Schwarz
ad7a104ab4 [#292] pruning: simplify table test driver 2020-09-02 22:45:44 +02:00
Christian Schwarz
639359f393 [#292] pruning: last_n: use snap name as fallback when creation is equal 2020-09-02 22:45:44 +02:00
Christian Schwarz
91e310b7e3 build: rpm + deb targets, build-in-docker targets, CircleCI pipeline rewrite
Co-authored-by: Armin Wehrfritz <dkxls23@gmail.com>
2020-09-02 21:34:52 +02:00
Christian Schwarz
5b30ad01ce make formatcheck: exit with non-zero status code if unformatted files are found
fixup of 7a5883d404
2020-09-02 21:33:36 +02:00
chenrui
8f4f9338a9
build: remove travis config
as I did not find it got referenced in the PR builds
2020-09-01 19:21:59 +02:00
Christian Schwarz
67bbce3c36
Merge pull request #375 from InsanePrawn/make_formatcheck
Add `make formatcheck` and reformat all files
2020-09-01 00:12:10 +02:00
InsanePrawn
180c3d9ae1 Reformat all files with make format.
Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-31 23:57:45 +02:00
InsanePrawn
7a5883d404 circleci: Add make formatcheck
Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-31 23:57:45 +02:00
InsanePrawn
c90acefacb Makefile: Add formatcheck target
- `make formatcheck` outputs the list of files that need formatting
  - `make format` applies the formatting and now also outputs the diffs.

Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-31 23:53:27 +02:00
InsanePrawn
d6799e08d8 lazy.sh: various fixes (#372)
- Emit warning and return 1 when no cmd is specified
- get docdep requirements from lazy.sh parent dir, not $GOPATH/src/zrepl/zrepl
- Explicitly warn when gopath is not in PATH
- Fix `release` cmd
- replace `! -z` with `-n`

Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-31 16:40:13 +02:00
Christian Schwarz
c44dccc34b build: circleci: switch to 'machine' executor for test-build-in-docker job
refs #356
2020-08-26 11:47:17 +02:00
Christian Schwarz
7f1695c457 docs: transport: fix easyrsa script (fixup of 6b4c6fc) 2020-08-23 20:36:43 +02:00
Christian Schwarz
e500d9ee26 package trace: track activeChildTasks in a set if debugEnabled=true
refs #358
2020-08-23 20:13:58 +02:00
Christian Schwarz
d5ce578929 [#358] platformtest: ReceiveForceIntoEncryptedErr: fix trace panic caused by dangling open zfs send process
```
panic: end task: 1 active child tasks: end task: task still has active child tasks

goroutine 1 [running]:
github.com/zrepl/zrepl/daemon/logging/trace.WithTask.func1.1(0xc000020680, 0xc000080f00)
	/home/cs/zrepl/zrepl/daemon/logging/trace/trace.go:221 +0x2f7
github.com/zrepl/zrepl/daemon/logging/trace.WithTask.func1()
	/home/cs/zrepl/zrepl/daemon/logging/trace/trace.go:237 +0x38
main.HarnessRun(0x7ffe0b49ad3a, 0x11, 0x7ffe0b49acf2, 0x1a, 0xc800000, 0x7ffe0b49ad19, 0x16, 0x1, 0x0, 0x0, ...)
	/home/cs/zrepl/zrepl/platformtest/harness/harness.go:168 +0xe54
main.main()
	/home/cs/zrepl/zrepl/platformtest/harness/harness.go:41 +0x2ef
```

fixes #358
2020-08-23 20:13:58 +02:00
Christian Schwarz
3f8fe3a368 [#346] do not rely on creation property for filesystem version diff
fixes #346
fixes #358
2020-08-23 19:21:55 +02:00
Christian Schwarz
6b4c6fc062 [#357] docs: update quickstart + tls transport to produce keypairs with subject alternative names
fixes #357
2020-08-22 03:05:30 +02:00
Christian Schwarz
e239d6f633 build: make platformtest-* usable
* use plain `go build` instead of the `go test` + `func TestMain` hack
* add target cover-platform and use that to generate coverage reports
2020-08-21 23:01:35 +02:00
InsanePrawn
0bbe2befce docs: prune: add prune interval visualisation
fixes #122

Co-Authored-By: Christian Schwarz <me@cschwarz.com>

Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-21 22:05:05 +02:00
InsanePrawn
fa4e048169 readme: fix typo
Signed-off-by: InsanePrawn <insane.prawny@gmail.com>
2020-08-21 22:05:05 +02:00
Christian Schwarz
4f9f21f7f2 logger: fix go-1.15-discovered conversion from int to string
logger/datastructures.go:85:10: conversion from Level (int) to string yields a string of one rune
2020-08-12 21:44:02 +02:00
Christian Schwarz
480176ba2d rpc/dataconn: fix go1.15-discovered recursive Error() method impl
rpc/dataconn/dataconn_client.go:69:9: Sprintf format %s with arg e causes recursive Error method call
2020-08-12 21:41:31 +02:00
Christian Schwarz
1190c0f6d2 docs: supporters: update 2020-08-12 21:38:23 +02:00
Christian Schwarz
720a284db5 dist/grafana: fix endpoint abstractions cache metric panel
fixup of 30cdc14
2020-08-04 01:36:31 +02:00
Hans Schulz
83fdffbcef replication: prometheus metric for number of failed replications in last attempt
- package replication: metric
- Grafana panel
- wiring
- changelog

Signed-off-by: Christian Schwarz <me@cschwarz.com>

closes #341
2020-08-04 01:19:44 +02:00
Christian Schwarz
0ee7a49d31 [#289] zfs: workaround for OpenZFS 0.7 dry send info with zero estimated size
fixes #289
2020-07-26 20:32:35 +02:00
Christian Schwarz
02db5994fe [#345] fix broken identification of parent-fs for initial replication ordering
fixup of 02807279

fixes #345
2020-07-26 20:32:35 +02:00
Christian Schwarz
1d7a84e8ae build: extract debian binary packaging workflow trigger into a reusable script 2020-07-26 20:32:35 +02:00
Christian Schwarz
a0e3dc7040 [#348] replication: add platformtest to check behavior on recv fail while still sending
Regression test for #348
2020-07-26 20:32:35 +02:00
Christian Schwarz
43495d70c7 [#348] replication: return errors if .Send rpc returns a nil SendRes
discovered while developing the next commit in this series
2020-07-26 20:32:35 +02:00
Christian Schwarz
fa586b493c [#348] fix crash on early-recv error
* The SendStream.Close() was not called by dataconn.Server, which left the zfs send process dangling.
* When the source job's ctx interceptor closed the task, the dangling zfs send was detect by the trace package and panicked.

    020-07-25T19:54:41-04:00 [ERRO][latitude][rpc.data][cyZj$J3Ca$J3Ca.CJwB]: cannot write send stream err="frameconn: shutting down"
    panic: end task: 1 active child tasks: end task: task still has active child tasks

    goroutine 196966 [running]:
    github.com/zrepl/zrepl/daemon/logging/trace.WithTask.func1.1(0xc000320680, 0xcde000)
            /home/jeremy/go/src/github.com/zrepl/zrepl/daemon/logging/trace/trace.go:221 +0x2f7
    github.com/zrepl/zrepl/daemon/logging/trace.WithTask.func1()
            /home/jeremy/go/src/github.com/zrepl/zrepl/daemon/logging/trace/trace.go:237 +0x38
    github.com/zrepl/zrepl/daemon/logging/trace.WithTaskAndSpan.func1()
            /home/jeremy/go/src/github.com/zrepl/zrepl/daemon/logging/trace/trace_convenience.go:41 +0x37
    github.com/zrepl/zrepl/daemon/job.(*PassiveSide).Run.func1(0xdcf780, 0xc0000a3560, 0xdc65a0, 0xc00035e620, 0xc0000a34d0)
            /home/jeremy/go/src/github.com/zrepl/zrepl/daemon/job/passive.go:194 +0x2e7
    github.com/zrepl/zrepl/rpc.NewServer.func3(0xdcf780, 0xc0001ce4b0, 0xdc65e0, 0xc00035e600, 0xc0000a34d0)
            /home/jeremy/go/src/github.com/zrepl/zrepl/rpc/rpc_server.go:82 +0xd5
    github.com/zrepl/zrepl/rpc/dataconn.(*Server).serveConn(0xc0000a2ba0, 0xc00018eca0)
            /home/jeremy/go/src/github.com/zrepl/zrepl/rpc/dataconn/dataconn_server.go:149 +0x3be
    github.com/zrepl/zrepl/rpc/dataconn.(*Server).Serve.func3(0xc0000b8180, 0xc0000a2ba0, 0xc00018eca0)
            /home/jeremy/go/src/github.com/zrepl/zrepl/rpc/dataconn/dataconn_server.go:108 +0x5d
    created by github.com/zrepl/zrepl/rpc/dataconn.(*Server).Serve
            /home/jeremy/go/src/github.com/zrepl/zrepl/rpc/dataconn/dataconn_server.go:106 +0x24a
    2020-07-25T19:58:55-04:00 [ERRO][latitude][rpc.data][Pt4F$gCWT$gCWT.fzhc]: cannot write send stream err="frameconn: shutting down"
    panic: end task: 1 active child tasks: end task: task still has active child tasks

fixes #348
2020-07-26 20:32:35 +02:00
Christian Schwarz
30cdc1430e replication + endpoint: replication guarantees: guarantee_{resumability,incremental,nothing}
This commit

- adds a configuration in which no step holds, replication cursors, etc. are created
- removes the send.step_holds.disable_incremental setting
- creates a new config option `replication` for active-side jobs
- adds the replication.protection.{initial,incremental} settings, each
  of which can have values
    - `guarantee_resumability`
    - `guarantee_incremental`
    - `guarantee_nothing`
  (refer to docs/configuration/replication.rst for semantics)

The `replication` config from an active side is sent to both endpoint.Sender and endpoint.Receiver
for each replication step. Sender and Receiver then act accordingly.

For `guarantee_incremental`, we add the new `tentative-replication-cursor` abstraction.
The necessity for that abstraction is outlined in https://github.com/zrepl/zrepl/issues/340.

fixes https://github.com/zrepl/zrepl/issues/340
2020-07-26 20:32:35 +02:00
Christian Schwarz
27673a23e9 config: add test for fromdefaults behavior 2020-07-26 20:32:35 +02:00
Christian Schwarz
95fc299733 daemon/job: test that sample configs are buildable 2020-07-26 20:32:35 +02:00
Christian Schwarz
4e702eedc9 cmd: zfs-abstraction list --json: fix panic
(was panicking because `abstractions` is in fact a channel
2020-07-26 20:32:35 +02:00
Christian Schwarz
8ff83f2f1a [#342] endpoint: always create unencrypted placeholder filesystems
This "breaks" the use case of receiving an unencrypted send into an encrypted receiver by setting the receiver's `root_fs`'s `encryption=on`.
"breaks" in air-quotes because we have not yet released a version of
zrepl with encrypted send support.

We will bring back the featured outlined above in a future release.
See https://github.com/zrepl/zrepl/issues/342#issuecomment-657231818 and following.
2020-07-26 20:32:35 +02:00
Christian Schwarz
4b8f0ad112 docs: supporters: update 2020-06-22 13:36:00 +02:00
Brian Candler
dbc8bbeb6a docs: config: prune: example: keep manual snapshots on receiver
Fixes #335
closes #336

Signed-off-by: Christian Schwarz <me@cschwarz.com>
2020-06-22 12:32:03 +02:00
Christian Schwarz
b3e856f40d docs: changelog: 0.3: fix broken issue link 2020-06-22 12:30:42 +02:00
Christian Schwarz
8e1937fe75 doc: fixup 0.3 changelog 05f1237a6d 2020-06-14 18:29:37 +02:00
Christian Schwarz
073514fc21 docs/publish.sh: only render latest (patch+rc) version for each (major,minor) versio. 2020-06-14 18:24:20 +02:00
Christian Schwarz
dab222d95f docs: GitHub Sponsors link 2020-06-14 15:26:05 +02:00
Christian Schwarz
a827894274 docs: add backup-to-external-disk quick-start guide and convert existing tutorial to quick-start guide
refs #219
fixes #329
2020-06-14 15:26:05 +02:00
Christian Schwarz
9a8d813d14 docs: fix typo in cli help for zfs-abstraction subcommand 2020-06-14 15:26:05 +02:00
Christian Schwarz
e391fa94f9 dist/grafana: update grafana dashboard
- uses version metric for 'instances up'
- displays active task count
- displays send abstractions cache entry count
- in general, graphs have a shorter y axis for better overview

fixes #332
2020-06-14 15:26:05 +02:00
Christian Schwarz
509185dfbe prometheus: expose zrepl version as const metric 2020-06-14 15:26:05 +02:00
Christian Schwarz
4b1b7a8561 envconst: queryable report of resolved variables + integration inot zrepl status --raw
fixes #299
refs #186
2020-06-14 15:26:05 +02:00
Christian Schwarz
b330ccca5d transport/ssh: bump go-netssh version to fix ssh client process leaks
fixes #322
2020-06-14 15:26:05 +02:00
Christian Schwarz
05f1237a6d docs: 0.3 changelog 2020-06-14 15:26:05 +02:00
Christian Schwarz
1c270b7e39 add option to disable step holds for incremental sends
This is a stop-gap solution until we re-write the pruner to support
rules for removing step holds.

Note that disabling step holds for incremental sends does not affect
zrepl's guarantee that incremental replication is always possible:

Suppose you yank the external drive during an incremental @from -> @to step:

* restarting that step or future incrementals @from -> @to_later` will be possible
  because the replication cursor bookmark points to @from until the step is complete
* resuming @from -> @to will work as long as the pruner on your internal pool doesn't come around to destroy @to.
    * in that case, the replication algorithm should determine that the resumable state
      on the receiving side isuseless because @to no longer exists on the sending side,
      and consequently clear it, and restart an incremental step @from -> @to_later

refs #288
2020-06-14 15:26:05 +02:00
Christian Schwarz
1b39e9d03c docs: update & extend replication overview wrt step holds + bookmarks 2020-06-14 15:21:36 +02:00