Commit Graph

38 Commits

Author SHA1 Message Date
Simone Gotti 5ab9f7c970 *: use etcd mutex TryLock
etcd PR 11104 (https://github.com/etcd-io/etcd/pull/11104) implemented mutex
TryLock. Since it's only available in etcd master just copy relevant code and
add a TODO to remove it when updating the etcd client to a version implementing
TryLock.

Use TryLock everywhere where it'll be useful.
2019-11-12 22:27:17 +01:00
Simone Gotti 1e70e3404b datamanager: skip already applied wals in writeDataSnapshot
As an optimization don't apply already applied wals.
2019-11-12 11:02:52 +01:00
Simone Gotti dfd0f8c806 datamanager tests: increase etcd waitdown timeout 2019-11-12 10:47:28 +01:00
Simone Gotti 72f279c4c3 *: improve error handling
* objectstorage: remove `types` package and move `ErrNotExist` in base package
* objectstorage: Implement .Is and add helper `IsErrNotExist` for `ErrNotExist`
* util: Rename `ErrNotFound` to `ErrNotExist`
* util: Add `IsErr*` helpers and use them in place of `errors.Is()`
* datamanager: add `ErrNoDataStatus` to report when there's not data status in ost
* runservice/common: remove `ErrNotExist` and use errors in util package
2019-11-11 12:17:35 +01:00
Simone Gotti 5af07d0852 objectstorage: use a single package
remove all the subpackages and just use a single package
2019-11-08 16:31:48 +01:00
Simone Gotti 35e1ec0e15 datamanager: remove old storage wals
Remove all wals not required by the existing data status files and not existing
in etcd.
2019-11-08 15:39:17 +01:00
Simone Gotti 9c0eb3d7ef datamanager: refactor ReadWal
make ReadWal directly return a *WalHeader
2019-11-08 13:24:43 +01:00
Simone Gotti 9c1f3b2a69 datamanager: check wal previouswalsequence is correct in initEtcd 2019-11-07 17:05:40 +01:00
Simone Gotti acd62a3f90 datamanager: don't create ost wal checkpointed files
currently creating .checkpointed files in the ost isn't useful.
We already have the data snapshot that reports the last checkpointed wal.
2019-11-07 10:32:41 +01:00
Simone Gotti 4fcb067052 datamanager: clean old data files
keep the last n (now set to 3) data status files and remove all other data status
files and unneeded data files.
2019-11-07 10:30:31 +01:00
Simone Gotti 52cb683267 datamanager: fix index creation on multiple data files
When during a checkpoint more than one file is created the entries position in
the index is not right since it's not reset at every new index.

Fix it and add related tests.
2019-11-06 15:33:40 +01:00
Simone Gotti 4b4416fc99 datamanager: add data sequence to data file name
When creating a datafile name make it start with the current data sequence. This
is useful in future to know which data sequence created a new data file.
2019-11-04 09:23:12 +01:00
Simone Gotti a7ca2848e6 datamanager: remove some logs from tests 2019-10-02 09:27:55 +02:00
Simone Gotti 6f3798e8fe *: use sleep timer in loops
So we'll react instantly to a context cancel instead of waiting on time.Sleep
returning.
2019-07-25 16:22:54 +02:00
Simone Gotti 3404cb94b9 datamanager: implement import/export
* export: exports the newest data checkpoint. It forces a checkpoint before
exporting (currently no wals are exported)

* import: cleans up etcd, creates a new datasnaphot from the provided import stream
and then initializes etcd. Currently no old data is removed from the object
storage but it's just ignored.
2019-07-25 11:12:49 +02:00
Simone Gotti 3987caf664 datamanager: implement maintenance mode
when datamanager is started in maintenance mode no goroutines are scheduled and
no initial etcd initialization is done
2019-07-24 17:37:27 +02:00
Simone Gotti f7175c4592 datamanager: save previous wal in waldata 2019-07-18 16:44:28 +02:00
Simone Gotti 18c5ae0492 datamanager: error if there're wals but not a datastatus in ost 2019-07-18 16:44:28 +02:00
Simone Gotti c034819087 datamanager: accept optional datastatus in initEtcd 2019-07-18 16:44:28 +02:00
Simone Gotti df716fccc6 datamanager: create a new wal and checkpoint in initEtcd
When doing an initEtcd (new instance or etcd reset) create a new wal (that will
have a new sequence epoch) and do a checkpoint.

In this way:

* readdb will detect that an epoch change and do a full resync
* we always have a data file (also if empty) that provides the last checkpointed
wal. This information could be used by readdb to resync
2019-07-18 16:44:28 +02:00
Simone Gotti cb2a871be6 datamanager: start initEtcd from last datastatus 2019-07-18 16:44:28 +02:00
Simone Gotti 445ef24daa datamanager: add option to force a checkpoint 2019-07-18 16:44:27 +02:00
Simone Gotti 512162bf98 datamanager: clean etcd data before reinitialization 2019-07-18 16:42:17 +02:00
Simone Gotti 8fbb9fdcbe datamanager: add deleteEtcd method 2019-07-18 16:40:28 +02:00
Simone Gotti c064e490b2 datamanager: add lock around compatchangegroups and initetcd
just to avoid concurrency errors when multiple instances are running
2019-07-10 10:20:03 +02:00
Simone Gotti d989fe9639 datamanager: always handle basepath
Currently we aren't setting a basepath and it wasn't always correctly handled.
Fix missing basepath handling and improve tests to also use a non empty
basepath.
2019-07-03 17:03:37 +02:00
Simone Gotti 49d0238a1e datamanager: fix linter errors
Fix errors reported by default golangci-lint linters
2019-07-02 14:27:51 +02:00
Simone Gotti 8d67844cc4 *: use vanity url
use agola.io domain
2019-07-01 11:40:20 +02:00
Simone Gotti 5e21089baa *: remove unneeded logging
remove many log.Info entries that where old debugging entries and move some of
them to the Debug level.
2019-06-14 11:28:00 +02:00
Simone Gotti 22bd181fc8 datamanager: implement data files splitting
split data files in multiple files of a max size (default 10Mib)
In this way every data snapshot will change only the datafiles that have some
changes instead of the whole single file.
2019-06-03 16:17:27 +02:00
Simone Gotti 9b2ce717c7 *: migrate to "golang.org/x/xerrors"
Just a raw replace of "github.com/pkg/errors".

Next steps will improve errors (like remote errors, api errors, not exist errors
etc...) to leverage its functionalities
2019-05-23 11:23:14 +02:00
Simone Gotti b3867fb7ca objectstorage: add posix standard storage
rename the previous posix storage to posixflat and make it currently not user
selectable (since I'm not sure it's really worth using it).

The new posix storage uses the filesystem without any escaping so it's not a
real flat namespace.

This isn't a real issue since also minio is not a flat namespace and we are so
forced to use it like a hierarchycal filesystem.
2019-05-21 15:17:53 +02:00
Simone Gotti bad18bf814 *: report objects size for objectstorage.WriteObject 2019-05-02 09:49:55 +02:00
Simone Gotti 34cfdfeb3b objectstorage: add size option to WriteObject
On s3 limit the max object size to 1GiB when the size is not provided (-1) or
the minio client will calculate a big part size since it tries to use the
maximum object size (5TiB) and will allocate a very big buffer in ram. Also
leave as commented out the previous hack that was firstly creating the file
locally to calculate the size and then put it (for future reference).
2019-05-02 09:47:38 +02:00
Simone Gotti e964aa3537 objectstorage: add persist option to WriteObject
This options is a noop on s3 but on the posix implementation it becomes useful
when there isn't the need to have a persistent file, thus avoiding some fsync
calls.
2019-05-01 15:06:47 +02:00
Simone Gotti 2634f7a3a0 datamanager: add function to get changegroups from etcd 2019-04-29 10:13:13 +02:00
Simone Gotti b85786dc56 datamanager: check changegroup name 2019-04-29 10:12:03 +02:00
Simone Gotti 2c3e6bf9e4 wal: update and rename to datamanager
* Rename to datamanager since it handles a complete "database" backed by an
objectstorage and etcd

* Don't write every single entry as a single file but group them in a single
file. In future improve this to split the data in multiple files of a max size.
2019-04-26 16:00:03 +02:00