We ignored this before, which meant that the DEVELOPER-mode check that we
delete the correct record didn't check that it wasn't already deleted.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We always know the length, so we don't need it. It causes much extra work
when we want to delete a record, which I suspect may cause issues amongst
some users who've been seeing gossip_store corruption.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rewriting the gossip_store is much more trivial when we don't have
any pointers into it, so add some simple offline compaction code
and disable the automatic compaction code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The crashes in #2750 are mostly caused by us trying to partially truncate
the store. The simplest fix for release is to discard the whole thing if
we detect a problem.
This is a workaround: it'd be far nicer to try to recover.
Fixes: #2750
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If something went wrong and there was an old one, we were
appending to it!
Reported-by: @SimonVrouwe
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We might have channel_announcements which have no channel_update: normally
these don't get written into the store until there is one, but if the
store was truncated it can happen. We then get upset on compaction, since
we don't have an in-memory representation of the channel_announcement.
Similarly, we leave the node_announcement pending until after that
channel_announcement, leading to a similar case.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We can't continue, since we've moved the indexes. We'll just crash
anyway, as seen from bugs #2742 and #2743.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I decided to try a faster implementation, only to find our crc32c was
not correct! Ouch.
I removed the crc32c functions from ccan/crc, and added a new crc32c
module which has the Mark Adler x86-64-optimized variants.
We bump gossip_store version again, since csums have changed.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
(Or, if we crashed before we got to write out the channel_update).
It's a corner case, but one reported by @darosior and reproduced
on my test node (both with bad gossip_store due to previous iterations
of this patchset!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This clarifies things a fair bit: we simply add and remove from the
gossip_store directly.
Before this series: (--disable-developer, -Og)
store_load_msec:20669-20902(20822.2+/-82)
vsz_kb:439704-439712(439706+/-3.2)
listnodes_sec:0.890000-1.000000(0.92+/-0.04)
listchannels_sec:11.960000-13.380000(12.576+/-0.49)
routing_sec:3.070000-5.970000(4.814+/-1.2)
peer_write_all_sec:28.490000-30.580000(29.532+/-0.78)
After: (--disable-developer, -Og)
store_load_msec:19722-20124(19921.6+/-1.4e+02)
vsz_kb:288320
listnodes_sec:0.860000-0.980000(0.912+/-0.056)
listchannels_sec:10.790000-12.260000(11.65+/-0.5)
routing_sec:2.540000-4.950000(4.262+/-0.88)
peer_write_all_sec:17.570000-19.500000(18.048+/-0.73)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We have a problem: if we get halfway through writing the compacted store
and run out of disk space, we've already changed half the indexes.
This changes it so we do nothing until writing is finished: then we
iterate through and update indexes. It also weans us off broadcast
ordering, which we can now eliminated.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We didn't count some records before, so we could compare the two counters.
This is much simpler, and avoids reliance on bs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
(We don't increment the gossip_store version, since there are only a
few commits since the last time we did this).
This lets the reader simply filter messages; this is especially nice since
the channel_announcement timestamp is *derived*, not in the actual message.
This also creates a 'struct gossip_hdr' which makes the code a bit
clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Keeping the uintmap ordering all the broadcastable messages is expensive:
130MB for the million-channels project. But now we delete obsolete entries
from the store, we can have the per-peer daemons simply read that sequentially
and stream the gossip itself.
This is the most primitive version, where all gossip is streamed;
successive patches will bring back proper handling of timestamp filtering
and initial_routing_sync.
We add a gossip_state field to track what's happening with our gossip
streaming: it's initialized in gossipd, and currently always set, but
once we handle timestamps the per-peer daemon may do it when the first
filter is sent.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
There's a corner case where otherwise a reader could see the header and
not the body of a message. It could handle that in various ways,
but simplest (and most efficient) is to avoid it happening.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We use the high bit of the length field: this way we can still check
that the checksums are valid on deleted fields.
Once this is done, serially reading the gossip_store file will result
in a complete, ordered, minimal gossip broadcast. Also, the horrible
corner case where we might try to delete things from the store during
load time is completely gone: we only load non-deleted things.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're about to bump version again, and the code to upgrade it was
quite hairy (and buggy!). It's not worthwhile for such a
poorly-tested path: I will just add code to limit how much incoming
gossip we get to avoid flooding when we upgrade, however.
I also use a modern gossip_store version in our test_gossip_store_load
test, instead of relying on the upgrade path.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
They're really gossipd-internal, and we don't want per-peer daemons
to confuse them with normal updates.
I don't bump the gossip_store version; that's coming with another update
anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Now we handle node_announcements properly, we have a failure case where we
try to move them when a channel is deleted while loading the store.
We're going to remove this soon, in favor of in-place delete, so
workaround this for now to avoid an assert() when we try to write to
the store while loading.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It turns out that we don't look at type when we return 0, but gcc isn't
quite smart enough for that. Initializing to -1 is good practice anyway
for the failure path.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Each destructor2 costs 40 bytes, and struct chan is only 120 bytes. So
this drops our memory usage quite a bit:
MCP bench results change:
-vsz_kb:580004-580016(580006+/-4.8)
+vsz_kb:533148
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We need to store the channel capacity for channel_announcement: hand it
in directly rather than having the gossip_store code do a lookup.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Save some overhead, plus gets us ready for giving subdaemons direct
store access. This is the first time we *upgrade* the gossip_store,
rather than just discarding.
The downside is that we need to add an extra message after each
channel_announcement, containing the channel capacity.
After:
store_load_msec:28337-30288(28975+/-7.4e+02)
vsz_kb:582304-582316(582306+/-4.8)
store_rewrite_sec:11.240000-11.800000(11.55+/-0.21)
listnodes_sec:1.800000-1.880000(1.84+/-0.028)
listchannels_sec:22.690000-26.260000(23.878+/-1.3)
routing_sec:2.280000-9.570000(6.842+/-2.8)
peer_write_all_sec:48.160000-51.480000(49.608+/-1.1)
Differences:
-vsz_kb:582320
+vsz_kb:582316
-listnodes_sec:2.100000-2.170000(2.118+/-0.026)
+listnodes_sec:1.800000-1.880000(1.84+/-0.028)
-peer_write_all_sec:51.600000-52.550000(52.188+/-0.34)
+peer_write_all_sec:48.160000-51.480000(49.608+/-1.1)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Entries we've already loaded expect to exist in the store. We could go
back and remove them all, but instead just truncate at the known-good
point.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Some tests require dev support, but the rest can run. We simplify
the gossip_store output so it's the same in non-dev mode too.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When we compact the store, we need to adjust the broadast index for
peers so they know where they're up to.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This requires some trickiness when we want to re-add unannounced channels
to the store after compaction, so we extract a common "copy_message" to
transfer from old store to new.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:36034-37853(37109.8+/-5.9e+02)
vsz_kb:577456
store_rewrite_sec:12.490000-13.250000(12.862+/-0.27)
listnodes_sec:1.250000-1.480000(1.364+/-0.09)
listchannels_sec:30.820000-31.480000(31.068+/-0.24)
routing_sec:26.940000-27.990000(27.616+/-0.39)
peer_write_all_sec:65.690000-68.600000(66.698+/-0.99)
MCP notable changes from previous patch (>1 stddev):
-vsz_kb:1202316
+vsz_kb:577456
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The next patch causes us to access the store while loading (we read
channel_updates for local peers), which messes up loading due to the
lseek involved.
Using pread() is atomic with seek & read, and also a bit more
efficient. Make the header contiguous too, while we're here.
We don't need pwrite: we always open with O_APPEND which means the
seek-to-end is implicit.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:36771-38289(37529.6+/-5.3e+02)
vsz_kb:1202316
store_rewrite_sec:12.460000-13.280000(12.784+/-0.29)
listnodes_sec:1.240000-1.410000(1.34+/-0.058)
listchannels_sec:29.850000-31.840000(30.908+/-0.69)
routing_sec:27.800000-31.790000(28.822+/-1.5)
peer_write_all_sec:66.200000-68.720000(67.44+/-0.84)
MCP notable changes from previous patch (>1 stddev):
-store_load_msec:39207-45089(41374.6+/-2.2e+03)
+store_load_msec:36771-38289(37529.6+/-5.3e+02)
-store_rewrite_sec:15.090000-16.790000(15.654+/-0.63)
+store_rewrite_sec:12.460000-13.280000(12.784+/-0.29)
-peer_write_all_sec:66.830000-76.850000(71.976+/-3.6)
+peer_write_all_sec:66.200000-68.720000(67.44+/-0.84)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When we no longer keep channel_updates in memory, there's a path where
we access them on load: when we promote a local channel to an
announced channel.
This breaks at the moment, since gs->fd == -1; change it to a writable
flag instead.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we need the payload, pull it from the gossip store.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:30189-52561(39416.4+/-8.8e+03)
vsz_kb:1812904
store_rewrite_sec:21.390000-27.070000(23.596+/-2.4)
listnodes_sec:1.120000-1.230000(1.176+/-0.044)
listchannels_sec:38.900000-50.580000(44.716+/-3.9)
routing_sec:45.080000-48.160000(46.814+/-1.1)
peer_write_all_sec:58.780000-87.150000(72.278+/-9.7)
MCP notable changes from previous patch (>1 stddev):
-vsz_kb:2288784
+vsz_kb:1812904
-store_rewrite_sec:38.060000-39.130000(38.426+/-0.39)
+store_rewrite_sec:21.390000-27.070000(23.596+/-2.4)
-listnodes_sec:0.750000-0.850000(0.794+/-0.042)
+listnodes_sec:1.120000-1.230000(1.176+/-0.044)
-listchannels_sec:30.740000-31.760000(31.096+/-0.35)
+listchannels_sec:38.900000-50.580000(44.716+/-3.9)
-routing_sec:29.600000-33.560000(30.472+/-1.5)
+routing_sec:45.080000-48.160000(46.814+/-1.1)
-peer_write_all_sec:49.220000-52.690000(50.892+/-1.3)
+peer_write_all_sec:58.780000-87.150000(72.278+/-9.7)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Instead of an arbitrary counter, we can use the file offset for our
partial ordering, removing a field. It takes some care when we compact
the store, however, as this field changes.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:34271-35283(34789.6+/-3.3e+02)
vsz_kb:2288784
store_rewrite_sec:38.060000-39.130000(38.426+/-0.39)
listnodes_sec:0.750000-0.850000(0.794+/-0.042)
listchannels_sec:30.740000-31.760000(31.096+/-0.35)
routing_sec:29.600000-33.560000(30.472+/-1.5)
peer_write_all_sec:49.220000-52.690000(50.892+/-1.3)
MCP notable changes from previous patch (>1 stddev):
-store_load_msec:35685-38538(37090.4+/-9.1e+02)
+store_load_msec:34271-35283(34789.6+/-3.3e+02)
-vsz_kb:2288768
+vsz_kb:2288784
-peer_write_all_sec:51.140000-58.350000(55.69+/-2.4)
+peer_write_all_sec:49.220000-52.690000(50.892+/-1.3)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is more compact, but also required once we replace the arbitrary
"index" with an actual offset into the gossip store. That will let us
remove the in-memory variants entirely.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:35685-38538(37090.4+/-9.1e+02)
vsz_kb:2288768
store_rewrite_sec:35.530000-41.230000(37.904+/-2.3)
listnodes_sec:0.720000-0.810000(0.762+/-0.041)
listchannels_sec:30.750000-35.990000(32.704+/-2)
routing_sec:29.570000-34.010000(31.374+/-1.8)
peer_write_all_sec:51.140000-58.350000(55.69+/-2.4)
MCP notable changes from previous patch (>1 stddev):
-vsz_kb:2621808
+vsz_kb:2288768
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
