This change fixes two things which can make reading from a corrupted
orders file inefficient.
* When a corrupted order is detected, but the underlying error is an
UnexpectedEOF (as opposed to a pb.Unmarshal error, for instance), there
is no point in attempting to read from the file another time to find an
additional uncorrupted order - we will continue to get UnexpectedEOF
errors until we seek to the very end of the file and get a normal EOF.
Instead, when UnexpectedEOF occurs, log and send metrics as with other
types of corruption, but do not attempt to read again.
* When a corrupted order is detected, instead of seeking forward only
one byte for the next attempt, seek forward by the size of entryHeader.
This cuts down on the number of iterations needed to find an uncorrupted
order after detecting a corrupted one.
Change-Id: Ie1a613127e29d29318584ec7f60e8f7554f73487
Missed one case of Unmarshal in the previous commit for V0 files (0f4e4969b7)
In V1, unmarshalling was being attempted before the checksum was
verified, so this commit moves those calls to the end of the V1 ReadOne
function.
Change-Id: Ic0b49f0bbc91fb61fb28af6003060994d0af22ed
In V0 orders files, unexpected EOF is correctly treated as a file
corruption, but pb.Unmarshal can also fail, and this is not treated as a
file corruption. This commit fixes that.
Change-Id: I6b446a10f4b1a5a44e832cbcc9bf8b2548cfcfeb
In production we are seeing ~115 storage nodes (out of ~6,500) are not using the new SettlementWithWindow endpoint (but they are upgraded to > v1.12).
We analyzed data being reported by monkit for the nodes who were above version 1.11 but were not successfully submitting orders to the new endpoint.
The nodes fell into a few categories:
1. Always fail to list orders from the db; never get to try sending orders from the filestore
2. Successfully list/send orders from the db; never get to calling satellite endpoint for submitting filestore orders
3. Successfully list/send orders from the db; successfully list filestore orders, but satellite endpoint fails (with "unauthenticated" drpc error)
The code change here add the following to address these issues:
- modify the query for ordersDB.listUnsentBySatellite so that we no longer select expired orders from the unsent_orders table
- always process any orders that are in the ordersDB and also any orders stored in the filestore
- add monkit monitoring to filestore.ListUnsentBySatellite so that we can see the failures/successes
Change-Id: I0b473e5d75252e7ab5fa6b5c204ed260ab5094ec
V1 allows the storagenode to continue reading orders from an
unsent/archived orders file, even if orders in the middle are corrupted.
Change-Id: Iea4117d55c05ceeb77f47d5c973e5ba95da46c66
Accidentally we imported the wrong monkit package with a previous
commit and made our go.mod and go.sum file unclean.
This should fix it.
Change-Id: I4c3c8b696f59cfd06dc2d5436bb7aea2805936ce
Since storage nodes check to see if any order files can be sent every 5
minutes, every storage node attempts to send orders to the satellite
within 5 minutes of each hour since this is when the files become
"available" to send. It is placing a lot of load on our satellite and
storage nodes are not being paid out properly due to timeouts during
order sending due to the increased satellite load.
Change-Id: I44d991b5884b8c11e8a3856d39aee8323f086b51
Abstract details of writing and reading data to/from orders files so
that adding V1 and future maintenance are easier.
Change-Id: I85f4a91761293de1a782e197bc9e09db228933c9
Combine store.writeLimit and store.writeOrder into
store.writeLimitAndOrder, which only requires a single call to
file.Write(). This simplifies code, but it also reduces the likelihood
of multiple calls to Write() increasing the likelihood of file
corruption.
Also combine the corresponding readLimit/readOrder functions for
consistency.
Change-Id: I62ed406fa2c02708465a678d18293f510f666440
We only need to lock aquire mutexes inside ListUnsentBySatellite when we
want to determine whether a file has an active enqueue in progress.
On some nodes, ListUnsentBySatellite can take a particularly long time, having
undesired side-effects, so if we can minimize locking time, those nodes
will be better off.
Also, lock archive mu during ListUnsentBySatellite so files cannot be
archived and listed at the same time.
Change-Id: Ieb7e2a759c20c724a74dd8315728c873ccab14a3
If we see an UnexpectedEOF error when attempting to read orders, return
the orders we have been able to read successfully and do not return an
error. This behavior ensures that the storagenode orders service
attempts to archive corrupted files and does not retry them repeatedly
and get stuck.
Change-Id: I0d00d1e174f968af6e99ca861eddad190f1339e2
satellite.DB.Console().Projects().GetAll database query
can be replaced with planet.Uplinks[0].Projects[0].ID
Change-Id: I73b82b91afb2dde7b690917345b798f9d81f6831
When we call ordersStore.BeginEnqueue, the unsent orders file for that
satellite and hour is prevented from being sent. It is freed when the
commit callback returned by BeginEnqueue is used. This change ensures
that we always call the commit callback, even when we have an empty
order or an order with Amount <= 0.
Change-Id: Ic4678f7eaa1e6957dd77d4bb5a23bb35d25b1e93
This change accomplishes multiple things:
1. Instead of having a max in flight time, which means
we effectively have a minimum bandwidth for uploads
and downloads, we keep track of what windows have
active requests happening in them.
2. We don't double check when we save the order to see if it
is too old: by then, it's too late. A malicious uplink
could just submit orders outside of the grace window and
receive all the data, but the node would just not commit
it, so the uplink gets free traffic. Because the endpoints
also check for the order being too old, this would be a
very tight race that depends on knowledge of the node system
clock, but best to not have the race exist. Instead, we piggy
back off of the in flight tracking and do the check when
we start to handle the order, and commit at the end.
3. Change the functions that send orders and list unsent
orders to accept a time at which that operation is
happening. This way, in tests, we can pretend we're
listing or sending far into the future after the windows
are available to send, rather than exposing test functions
to modify internal state about the grace period to get
the desired effect. This brings tests closer to actual
usage in production.
4. Change the calculation for if an order is allowed to be
enqueued due to the grace period to just look at the
order creation time, rather than some computation involving
the window it will be in. In this way, you can easily
answer the question of "will this order be accepted?" by
asking "is it older than X?" where X is the grace period.
5. Increases the frequency we check to send up orders to once
every 5 minutes instead of once every hour because we already
have hour-long buffering due to the windows. This decreases
the maximum latency that an order will be reported back to
the satellite by 55 minutes.
Change-Id: Ie08b90d139d45ee89b82347e191a2f8db1b88036
* Add all new orders to the orders filestore instead of the database.
* Submit orders from the filestore to the new satellite SettleWindow
endpoint.
The orders filestore will eventually replace the orders DB completely.
For now, we will still be checking the orders DB and submitting those
orders if they exist. In a later release, we will completely remove the
orders DB, but we need both the DB and filestore for the transitionary
period.
Change-Id: Iac8780fd5ab770296181bbd313e1d335f072d4dc
* Instead of archiving a list of orders and deleting an "unsent" file in
separate steps, archival simply moves the old unsent file to a new
archived file
* Add maxInFlightTime to be used along with grace period for sending
buffer
* Create unsent/archival directories in constructor
* Code cleanup
Change-Id: Ia3bc2aaf60cced6c6d413465423d78c7d5151188
* Allow orders to be archived after being settled successfully with the
satellite.
* Allow for cleanup of orders that were archived before a certain time.
* Rewrite other parts of the orders file store to work better with new
design.
Change-Id: I39bea96d80e66a324ec522745169bd6d8b351751
* Will replace order limits database.
* This change adds functionality for storing and listing unsent orders.
* The next change will add functionality for order archival after
submission.
Change-Id: Ic5e2abc63991513245b6851a968ff2f2e18ce48d
Most places now need the NodeURL rather than the ID and Address
separately. This simplifies code in multiple places.
Change-Id: I52621d8ca52296a8b5bf7afbc1001cf8bfb44239
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
When error is formatted using %v it's not possible to check
whether the error was caused by a context cancellation.
Change-Id: Ia77dfb0817e49d9a7b168c12a6300d131007d0ee
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
all of the packages and tests work with both grpc and
drpc. we'll probably need to do some jenkins pipelines
to run the tests with drpc as well.
most of the changes are really due to a bit of cleanup
of the pkg/transport.Client api into an rpc.Dialer in
the spirit of a net.Dialer. now that we don't need
observers, we can pass around stateless configuration
to everything rather than stateful things that issue
observations. it also adds a DialAddressID for the
case where we don't have a pb.Node, but we do have an
address and want to assert some ID. this happened
pretty frequently, and now there's no more weird
contortions creating custom tls options, etc.
a lot of the other changes are being consistent/using
the abstractions in the rpc package to do rpc style
things like finding peer information, or checking
status codes.
Change-Id: Ief62875e21d80a21b3c56a5a37f45887679f9412
