Curently, storage nodes only report their capacity to satellites
once per hour. If a node fills up, it will fail all uploads until
the next contact cycle begins. With these changes, at the end of an
upload we check whether the MinimumDiskSpace threshold has been
passed. If so, trigger the monitor chore to update the node's
capacity, then trigger the contact chore to report the new
capacity to the satellites
Change-Id: Ie6aadaade1e2c12c87e03f8ff9059a50121380a0
this commit updates our monkit dependency to the v3 version where
it outputs in an influx style. this makes discovery much easier
as many tools are built to look at it this way.
graphite and rothko will suffer some due to no longer being a tree
based on dots. hopefully time will exist to update rothko to
index based on the new metric format.
it adds an influx output for the statreceiver so that we can
write to influxdb v1 or v2 directly.
Change-Id: Iae9f9494a6d29cfbd1f932a5e71a891b490415ff
This test checks that we are actually walking over the pieces when
starting the cache, and that it is returning expected values.
A recent outage was partially caused by the fact that this cache was
accidentally reading itself (via the pieces store, which has the cache
embedded). This test ensures that does not happen, and checks that when
the cache's `Run` method is called, the space used values are read from
disk and accurately update the cache.
Change-Id: I9ec61c4299ed06c90f79b17de3ffdbbb06bc502e
As a workaround it was set to 0 in previous release. Now according to the TOC must be set to 500GB.
Change-Id: Ia2743d49e86683396958aff51b95df743af4f872
http.FileServer relies on mime types defined in the operating system.
These values may be misconfigured, so a javascript file might
end up being served as "plain/text".
Change-Id: I3c13c8a9ac484bd765a4de0f8253bfe40dde7513
it was noticed that if you had a long lived transaction A that
was blocking some other transaction B and A was being aborted
due to retriable errors, then transaction B was never given
priority. this was due to using savepoints to do lightweight
retries.
this behavior was problematic becaue we had some queries blocked
for over 16 hours, so this commit addresses the issue with two
prongs:
1. bound the amount of time we will retry a transaction
2. create new transactions when a retry is needed
the first ensures that we never wait for 16 hours, and the value
chosen is 10 minutes. that should be long enough for an ample
amount of retries for small queries, and huge queries probably
shouldn't be retried, even if possible: it's more preferrable to
find a way to make them smaller.
the second ensures that even in the case of retries, queries that
are blocked on the aborted transaction gain priority to run.
between those two changes, the maximum stall time due to retries
should be bounded to around 10 minutes.
Change-Id: Icf898501ef505a89738820a3fae2580988f9f5f4
instead of aborting on the first error, so that we can hit all
satellites and get the best numbers we can
Change-Id: I21d5163884940612d7d39eaf73a6fac07235cd9e
We have added a bug with v0.31.7 and deploying it would kick out all the
storage nodes that are full. Easy fix is setting the requirment to 0.
That will allow them to still start up even if they are full.
Change-Id: Ie66f369952d929fcfd47f44f6e5e57eea8f51ff6
Per default our server address is listening on all IP addresses on the machine.
This caused our preflight check to fail, as it did not have an hostname to lookup.
With this change, we are fine with this and go ahead.
Change-Id: I9eb5c891c099eb35f679d6d7e79ec38bb43b619f
1 transfer with a minimum speed of 128 Bytes was a nice try but it is
way too low. Even a pi3 was able to handle 7 grpc transfers. We have 4
satellites and with 5 concurrent transfers that should be a total of 20
concurrent transfers. Each transfer will have a minimum speed of 5KB/s.
That should give us a better througput and still be Ok on a pi3.
Change-Id: I650a7baf890080901ef70ea3b5636d93009b4e60
With the v0.30.5 release we asked the storage node operators to manually
enable the preflight check while they are in front of their machine. We
didn't want to risk taking too many storage nodes offline at the same
time because of some unknow bug. The preflight check worked. We have no
negative feedback. We can now enable it by default.
Change-Id: Ic670ee52becd0b35eca84af7a0841ea983d7b19d
clock sync
Change 24h and 1h to 30m and 10m respectively for clock sync. If a
storagenode's clock is off by more than 30m for every trusted satellite,
it will not start. If it is off by more than 10m for any trusted
satellite, a warning is displayed.
Change-Id: I05ef611a30a49c1783e3b68b513745922c2f7e28
this is to help protect against intentional or unintentional
slowloris style problems where a client keeps a tcp connection
alive but never sends any data. because grpc is great, we have
to spawn a separate goroutine for every read/write to the stream
so that we can return from the server handler to cancel it if
necessary. yep. really.
additionally, we update the rpcstatus package to do some stack
trace capture and add a Wrap method for the times where we want
to just use the existing error.
also fixes a number of TODOs where we attach status codes to the
returned errors in the endpoints.
Change-Id: Id8bb8ff84aa34e0f711b0cf9bce3908b36a1d3c1
A few variables were not renamed to the new standard piecesTotal and
piecesContentSize, so it was unclear which value was being used. These
have been updated, and some comments made more thorough.
Change-Id: I363bad4dec2a8e5c54d22c3c4cd85fc3d2b3096c
This change updates the storagenode piecestore apis to expose access to
the full piece size stored on disk. Previously we only had access to
(and only kept a cache of) the content size used for all pieces. This
was inaccurate when reporting the amount of disk space used by nodes.
We now have access to the total content size, as well as the total disk
usage, of all pieces. The pieces cache also keeps a cache of the total
piece size along with the content size.
Change-Id: I4fffe7e1257e04c46021a2e37c5adc6fe69bee55
With this change RS configuration will be set on satellite. Uplink with
get RS values with BeginObject request and will use it. For backward
compatibility and to avoid super large change redundancy scheme stored
with bucket is not touched. This can be done in future.
Change-Id: Ia5f76fc10c37e2c44e4f7b8754f28eafe1f97eff
Replace all the remaining uses of sql.DB with tagsql.DB to
fix issues with context cancellation.
Introduce tagsql.Open which helps to get rid of all tagsql.Wrap-s.
Use tagsql in cockroachkv and postgreskv.
Change-Id: I8946d203341cb85a25976896fc7881e1f704e779
Migration step was closing a database that was used by
the migration itself. There is an active tranasction
over the database.
Instead of closing in the same transaction we can wait
until restart for the database cleanup.
Change-Id: Ic971d8cea81a3ab783f4a1bdc6357009c8b31386
Also added temporary types withRebind and withTagTx,
which will be later removed. Currently they help to avoid
changing the whole codebase at the same time.
Change-Id: I7f07ba8f4709a23a463bfa67464628665a05808f
storagenode database preflight check.
Disable preflight database check by default, and have the option to
enable it. This will allow us to enable it once it is definitely
working.
Also change the name of the config flag for preflight time sync.
Change-Id: Ie2e20f9e25dcb38794eafa7e1505e7c6ff287c99
On pieces usage cache init we now load the trash info from the db. Also
fixes a test that was masking the failure here.
Change-Id: I9ff7da5bc6c0f74cf0942e20931b40e0c88d70fa
for storagenode
Ensure that database schema matches latest test migration schema before
allowing the node to start up.
Ensure minimal read/write functionality for each storagenode database
before allowing the node to start up.
This will eliminate many unhandled audit errors we are seeing.
Change-Id: Ic0e628b04a9c35b7a8243f6a81d4683918170ba9
This reverts commit 8e242cd012.
Revert because lib/pq has known issues with context cancellation.
These issues need to be resolved before these changes can be merged.
Change-Id: I160af51dbc2d67c5449aafa406a403e5367bb555
this will allow for some nice runtime analysis down the road.
also, this allows for wrapping database handles in a way that
can interact with these contexts
requires https://review.dev.storj.io/c/storj/dbx/+/514
Change-Id: Ib087b7cd73296dd2c1e0331314da34d861f61d2b
When error is formatted using %v it's not possible to check
whether the error was caused by a context cancellation.
Change-Id: Ia77dfb0817e49d9a7b168c12a6300d131007d0ee
This code needs to work against cockroachDB, so transactions must be retried
when a retryable error is returned. This change puts migrate
transactions into the dbutil.WithTx transactional helpers to achieve
this in the easiest way.
Change-Id: Ib930e82d55cb0257357a222ce9131e6e53372c03
Remove starting up messages from peers. We expect all of them to start,
if they don't, then they should return an error why they don't start.
The only informative message is when a service is disabled.
When doing initial database setup then each migration step isn't
informative, hence print only a single line with the final version.
Also use shorter log scopes.
Change-Id: Ic8b61411df2eeae2a36d600a0c2fbc97a84a5b93
Remove direct dependency on uplink.RSConfig, this simplifies
moving the config file without introducing weird dependencies.
Change-Id: I7fd2a145401e0205d7047631df9d2810241efeec
This commit adds functionality to include the space used in the trash
directory when calculating available space on the node.
It also includes this trash value in the space used cache, with methods
to keep the cache up-to-date as files are trashed, restored, and
emptied.
As part of the commit, the RestoreTrash and EmptyTrash methods have
slightly changed signatures. RestoreTrash now also returns the keys that
were restored, while EmptyTrash also returns the total disk space
recovered. Each of these changes makes it possible to keep the cache
up-to-date and know how much space is being used/recovered.
Also changed is the signature of PieceStoreAccess.ContentSize method.
Previously this method returns only the content size of the blob,
removing the size of any header data. This method has been renamed
`Size` and returns both the full disk size and content size of the blob.
This allows us to only stat the file once, and in some instances (i.e.
cache) knowing the full file size is useful.
Note: This commit simply adds the trash size data to the piece size data
we were already collecting. The piece size data is not accurate for all
use-cases (e.g. because it does not contain piece header data); however,
this commit does not fix that problem. Now that the ContentSize (Size)
method returns the full size of the file, it should be easier to fix
this problem in a future commit.
Change-Id: I4a6cae09e262c8452a618116d1dc66b687f59f85
Fixes a data race caused by not waiting for workers to finish
before shutting down. Currently this ended up failing logging
because it was closed when test tried to write to it.
Change-Id: I074045cd83bbf49e658f51353aa7901e9a5d074b
- also updated ping chore to pick up trust changes
- fixed small typo in blueprint
- fixed flags for storj-sim
- wired up changes to testplanet
Change-Id: I02982f3a63a1b4150b82a009ee126b25ed51917d
planet.Start starts a testplanet system, whereas planet.Run starts a testplanet
and runs a test against it with each DB backend (cockroach compat).
Change-Id: I39c9da26d9619ee69a2b718d24ab00271f9e9bc2
Satellite URL is a stricter form of the STORJ Node URL. It requires both
the ID and port specifier.
Change-Id: I7fd302064f864c1de8240a7915bf5263b898dfd1
* pkg/pg: Add new service function storage node
Add a new service function to the storage node piece store for deleting
pieces when satellites request them.
* storagenode/piecestore: Add endpoint to delete piece
Add a new endpoint to receive from trusted satellites to delete a piece.
* private/testplanet: Fix storagenode mock
Add to the storagenode mock the new endpoint method.
* proto.lock: Update it with the last protbuff changes
* storagenode/piecestore: Reuse test piece upload
Extract the repeated logic from several tests functions for uploading a
test piece to a test helper function.
* uplink/piecestore: Implement client side method
Implement the client side method of the new piecestore RPC function.
* storagenode/piecestore: Add test DeletePiece endpoint
Implement a test for the DeletePiece new endpoint method.
* put TestCreateV0 back in StoreForTest
* avoid direct handles to V0 pieceinfo db
* type mismatch fix
* use storage.Blobs interface in store_test.go
..instead of filestore.Store. this will allow filestore.Store to become
unexported.
* unexport filestore.Store
rename it to blobStore. things should use the storage.Blobs interface
instead. changes in this commit are purely mechanical (made through the
"refactor" tool in Gocode followed by search/replace on the word "Store"
within the storage/filestore/ directory).
* kill filestore.StoreForTest
now that filestore.blobStore is unexported, there isn't a need for a
specialized wrapper type. this (not coincidentally) also makes it
possible for the WriterForFormatVersion() method on
storagenode/pieces.StoreForTest to work, without requiring everything to
wrap the store.blobs attribute in a filestore.StoreForTest, which was
impractical.
* change satellite.Peer name to Core
* change to Core in testplanet
* missed a few places
* keep shared stuff in peer.go to stay consistent with storj/docs
* improve errors in satellite contact endpoints
* add changes per CR comments
* update pingback method so it still updates node table
* fix err and returns
* fix zap logging to be better
When the contact chore starts running before the monitor service has
provided any useful capacity data, the first outgoing contact has
not-very-helpful data for the satellite. This change causes the contact
chore to wait until capacity data is available. The wait should be quite
short in all reasonable cases: even when a node starts with a lot of
stored pieces and no cached spaceUsedDB data, new data will have been
calculated and cached by the call to
`peer.Storage2.CacheService.Init(ctx)` in `storagenode.cmdRun()` before
`peer.Run(ctx)`.
Change-Id: Ibc26d5c1fc10a23006c00bc3f13ff6cf71f8bf1d
* Make the exiting node check piece hashes, piece IDs, and piece hash signatures before relaying successful transfer data to the satellite.
* Enable immediate graceful exit failure for "successful" transfers that fail satellite-side validation.
* Move transfer piece logic in storagenode worker to separate function (to make the worker easier to understand)
* add overall failure percentage check and inactive time frame check before sending a response to sno
* update comment
* delete node from transfer queue if it has been inactive for too long
* fix linting error
* add test config value
* fix nil pointer
* add config value into testplanet
* add unit test for overall failure threshold
* move timeframe threshold to chore
* update protolock
* add chore test
* add per peiece failure count logic
* change config name from EndpointMaxFailures to MaxFailuresPerPiece
* address comments
* fix linting error
* add error handling for no row returned from progress table
* fix test for graceful exit chore on storagenode
* fix typo InActive -> Inactive
* improve readability for failure threshold calculation
* update config lock
* change error handling for GetProgress in graceful exit endpoint on the satellite side
* return proper rpc error in endpoint
* add check in chore test for checking finish timestamp and queue
libuplink was incorrectly setting timeouts to 10 seconds still, but
should have been at least 10 minutes. the order sender was setting them
to 1 hour. we don't want timeouts in uplink-side logic as it establishes
a minimum rate on tcp streams.
instead of all of this, just use tcp keep alive. tcp keep alive packets are
sent every 15 seconds and if the peer stops responding the connection
dies. this is enabled by default with go. this will kill tcp connections
when they stop working.
Change-Id: I3d7ad49f71950b3eb43044eedf4b17993116045b
The upload code currently updates the usage in a deferred call to saveOrder().
The consequence is that in the success case, the RPC is completed before
the usage has been updated.
This change repurposes the deferred call to update usage in the
failure case, while explicitly updating the usage before completing the
RPC.
This fixes some test flakiness when using dRPC. gRPC waits until the final status is written before a Recv call completes, and the final status is written by the server after the handler function has exited. In practice this means that the client is blocked until the defer call is also finished. So this change will not change performance at all.
It has two advantages:
(1) It fixes test flakiness
and, more importantly:
(2) reduces the chances that someone will accidentally write a flaky test in the future
* add exit-status command
* remove todo and fix format
* fix status display
* change startExit to exit progress
* fix linting error
* add successful column in exit progress
* fix test
* remove extra new line
* fix TYPOS
* format the percentage better
This change adds a trusted registry (via the source code) of node address to node id mappings (currently only for well known Satellites) to defeat MITM attacks to Satellites. It also extends the uplink UI such that when entering a satellite address by hand, a node id prefix can also be added to defeat MITM attacks with unknown satellites.
When running uplink setup, satellite addresses can now be of the form 12EayRS2V1k@us-central-1.tardigrade.io (not even using a full node id) to ensure that the peer contacted is the peer that was expected. When using a known satellite address, the known node ids are used if no override is provided.
