Static Channel Backup. #5288

adi2011 · 2022-05-24T21:09:08Z

SCB will enable users to backup channels in an encrypted file emergency.recover, and recover funds upon data corruption. I've added some RPCs and modifications to the in-memory channel struct to enable this.

-staticbackup: Returns SCB (in the form of a TLV) of all the existing channels.

-recoverchannel: Commits channels (with enough data to recover funds and the rest is dummy data) to the DB.

-makesecret: Gives out pseudorandom keys (based on the hex input) derived from the HSM.

-emergencyrecover: Recover all the stored channels in case of data loss

There's a new TLV message as well (scb_wire) which is used to serialize/deserialize the data for a channel.
The scb file is maintained by a plugin.

Thanks, @niftynei, and @rustyrussell for all the valuable inputs and guidance. ❤️

cdecker

Excellent work @adi2011, I saw your PR and got excited to take a look. I know this is just a draft, but it's already looking very good.

Since I was here I added a couple of comments, mostly unimportant stuff like formatting (hadn't realized the Draft status) so feel free to ignore them as you likely would have addressed them before undrafting.

One bikeshedding topic I'd like to tackle though is the name SCB, as it always annoyed me that it is being positioned as a backup. Would you be open to call it "emergency recovery" or something like that?

cdecker · 2022-05-25T10:07:05Z

lightningd/hsm_control.c

@@ -146,6 +147,43 @@ static struct command_result *json_getsharedsecret(struct command *cmd,
 	return command_success(cmd, response);
 }

+static struct command_result *json_getsecret(struct command *cmd,


Is this not already achieved by getsharedsecret? I'm always a bit hesitant to add generic "give me a secret" methods as they may lead to extracting secrets we don't want to give away.

getsharedsecret needs a compressed public key and then it performs ecdh() to give out a key for connection handshake and for onion wrapping.

We needed a pseudorandom key derived from hsm_secret, using hkdf_sha256() taking in a string as info, but I get your point.
what if instead of taking in an 'info' string from the user I hard code it to "peer backup" to get a specific key used to encrypt the scb file? and rename it to getbackupsecret?

Ummmm just generate a random compressed public key? Pick one out of a hat, then call getsharedsecret and then use the result as the encryption key, AND store the compressed public key in the first 33 bytes of the emergency-recovery file. This lets you recover it anyway with the same hsm_secret; you only need the session pubkey.

Secondarily, the VLS people have implemented a hardware version of the hsmd, and adding a new API to the hsmd requires that we at least inform them so they can implement as well. So I would be reluctant to add a new hsmd_derive_secret message here, we should try to use getsharedsecret which they already implemented (hsmd_ecdh_req, hsmd_ecdh_resp) and just give a random point that you store with the recovery file anyway. The random point will not leak the encryption key because of ECDH magic so it is safe to put it right beside the encrypted blob in the recovery file.

I've modified it by deriving a key from the HSM and then derived the keys for this RPC from that one, I don't think that'd cause any security issue and this can be used for generating keys for other things as well in the future.

cdecker · 2022-05-25T10:10:35Z

lightningd/channel.c

+	}
+	else{


Nit: whitespace -> } else {

cdecker · 2022-05-25T10:10:57Z

lightningd/channel.c

-	channel->last_tx = tal_steal(channel, last_tx);
-	channel->last_tx->chainparams = chainparams;
-	channel->last_tx_type = TX_UNKNOWN;
+	if(last_tx){


Nit: whitespace: if (last_tx) {

cdecker · 2022-05-25T10:15:03Z

lightningd/peer_control.c

+}
+
+static const struct json_command hex_scb_command = {
+	"hex-scb",


Please don't use - in command names, they translate very poorly into other languages, requiring workarounds and renaming. We usually just go with something like hexscb. Also it'd be nicer to have staticbackup as a name instead, hex just being the format, allowing us to iterate in the future, without requiring a rename.

My bad, Yes I'll change it to staticbackup...

Maybe static_backup is better? it makes the docs more accessible for screen readers too.

But usually, the _ will translate directly into other languages with the camel case, so it is more difficult for single word command name.

But it is minor if it requires a lot of work!

cdecker · 2022-05-25T10:16:53Z

lightningd/channel.c

@@ -756,6 +756,14 @@ void channel_fail_permanent(struct channel *channel,
 			    const char *fmt,
 			    ...)
 {
+	/* FIXME: should I compare the whole scid for stub channel?
+		I doubt there are actually any channels in the Genesis block(1x1x1) */
+	if(channel->scid->u64 >> 40 == 1){


Is 1x1x1 a special value in this case? We're not actually guaranteed to have an scid at the time we create the recovery data, since the channel may not have yet confirmed.

I see it below, seems a bit strange to use this, and not a way more unreasonable UINT64_MAX, but it'd ultimately be the same anyway. Maybe it'd be better to just keep it NULL?

I am setting it to 1x1x1 to indicate (while loading the in-memory channel from DB) that it's a stub channel and we need to send the error to the peer, nothing else. It was on my to-do list to set it to 0x0x0 (as Rusty suggested), but it won't make any difference.

I think setting it to NULL will create confusion with the unconfirmed channels which also don't have any scid? If that's not the case I am happy to set it to NULL.

I like the idea to make the code unambiguous for the new readers, maybe we can make a #DEFINE GENESIS_SHORT_ID "1x1x1"

P.S: I'm assuming that it is easy to convert into a string the channel->scid

I need to remember to read the PR from bottom to top, the Rusty suggestion is better I think #5288 (comment)

lightningd/opening_control.c

cdecker · 2022-05-25T10:22:23Z

lightningd/opening_control.c

+			      AMOUNT_SAT(500000000), // FIXME: Is this correct? 
+				  /*Need to keep this very high from trimming the fees 
+				  	in onchaind.c(main).*/


Might be worthwile to add the actual funding amount to the recovery data? It's not going to change anyway, so it'd be safe.

Yeah, I thought of adding it, but then, these channels will be unilaterally closed by the peer and the wallet will forget them after sweeping the funds. So I don't think there's any point in swelling up the SCB file with unnecessary info.

How about setting it to an unrealistically high number?

2,100,000,000,000,001 satoshis?

2,100,000,000,000,001 satoshis?

Fixed in later commits.

cdecker · 2022-05-25T10:23:49Z

lightningd/opening_control.c

@@ -1227,3 +1406,13 @@ static const struct json_command fundchannel_complete_command = {
    "with {psbt}. Returns true on success, false otherwise."
 };
 AUTODATA(json_command, &fundchannel_complete_command);
+
+static const struct json_command json_commitchan_command = {
+    "emergencystubchannel",


recoverchannel?

Sure, I am anyways bad at naming things in general...

cdecker · 2022-05-25T10:24:48Z

wallet/wallet.c

+			that can motivate him to cheat.*/
+		chan->error = towire_errorfmt(w->ld, 
+									  &cid,
+									 "We can't be together anymore.."


I like this error message 👍

cdecker · 2022-05-25T10:28:09Z

plugins/chanbackup.c

+}
+
+/* checks if the SCB file exists, creates a new one in case it doesn't. */
+static void maybe_create_new_scb(struct plugin *p, const char *scb_buf)


This overwrites the file in place, which does not guarantee atomic operations. It's better to create a temporary file recover.tmp, write to that and then rename recover.tmp to recover. That guarantees that a partial write (plugin crash, loss of power, etc) doesn't corrupt your recovery file.

Yes, that was my doubt as well. I'll do this right away 🙌, Thanks!

Required reading: UPDATE IN PLACE: A Poison Apple?

adi2011 · 2022-05-25T18:42:21Z

Excellent work @adi2011, I saw your PR and got excited to take a look. I know this is just a draft, but it's already looking very good.

Thank you so much for taking out the time to check this out. I'll soon remove the draft tag after testing notification responses of the plugin, looking forward to more valuable feedback 😄

One bikeshedding topic I'd like to tackle though is the name SCB, as it always annoyed me that it is being positioned as a backup. Would you be open to call it "emergency recovery" or something like that?

Yes, I agree with this. Do I need to change the names of the TLV message and commands as well?

rustyrussell

OK, didn't review plugin yet! But lots of comments on the other stuff. Mostly easy things, though!

This is an excellent first PR :)

hsmd/libhsmd.c

lightningd/hsm_control.c

lightningd/channel.c

lightningd/opening_control.c

wallet/wallet.c

adi2011 · 2022-05-25T20:20:30Z

Thank you so much for the review and sorry for the NIT:s (I'll keep them in mind from now onwards), I'll try to incorporate all the changes that you and @cdecker suggested by the end of this week.

adi2011 · 2022-05-27T20:51:21Z

I hope after sorting out the final commit's FIXME, this would be mergeable 🤞🏻 ...

ZmnSCPxj · 2022-05-31T22:59:56Z

Suggest adding a section to doc/BACKUP.md about this. In particular, since this now uses an atomic construct-temporary-file-and-rename, it is very safe for an automated backup process to hardlink the recovery file, copy the hardlinked name, then unlink, and the automated backup process assuredly will never acquire a corrupted copy, either.

vincenzopalazzo

I make the first review of this, so I can continue later.

basically are only C-style code fixes, and some json rpc command good manners when there is a failure.

vincenzopalazzo · 2022-05-31T22:47:53Z

lightningd/Makefile

@@ -90,6 +90,7 @@ LIGHTNINGD_COMMON_OBJS :=			\
 	common/features.o			\
 	common/fee_states.o			\
 	common/peer_status_wiregen.o		\
+	common/scb_wiregen.o            \


Suggested change

common/scb_wiregen.o \

common/scb_wiregen.o \

we can alienate the \

vincenzopalazzo · 2022-05-31T22:48:20Z

lightningd/channel.c

+	} else {
+		channel->last_tx = NULL;
+	}


Suggested change

} else {

channel->last_tx = NULL;

}

} else

channel->last_tx = NULL;

vincenzopalazzo · 2022-05-31T22:48:37Z

lightningd/channel.c

+		channel->error = towire_errorfmt(peer->ld, 
+										&channel->cid,
+										"We can't be together anymore."
+										);


This is unformatted?

vincenzopalazzo · 2022-05-31T22:48:53Z

lightningd/channel.c

+	if (is_stub_scid(channel->scid)) {
+		return; 
+	}


Suggested change

if (is_stub_scid(channel->scid)) {

return;

}

if (is_stub_scid(channel->scid))

return;

vincenzopalazzo · 2022-05-31T22:50:22Z

lightningd/hsm_control.c

+		   NULL))
+		return command_param_failed();
+
+	u8 *point = tal_dup_arr(cmd, u8, (u8 *)info, strlen(info), 0);


Suggested change

u8 *point = tal_dup_arr(cmd, u8, (u8 *)info, strlen(info), 0);

u8 *point = tal_dup_arr(cmd, u8, (u8*)info, strlen(info), 0);

vincenzopalazzo · 2022-05-31T23:02:50Z

plugins/chanbackup.c

+	if (fd < 0) {
+		/* If this is not the first time we've run, it will exist. */
+		if (errno == EEXIST)
+			return;
+	}


Suggested change

if (fd < 0) {

/* If this is not the first time we've run, it will exist. */

if (errno == EEXIST)

return;

}

if (fd < 0)

/* If this is not the first time we've run, it will exist. */

if (errno == EEXIST)

return;

vincenzopalazzo · 2022-05-31T23:03:13Z

plugins/chanbackup.c

+	if (fsync(fd) != 0) {
+		unlink_noerr("scb.tmp");
+	}


Suggested change

if (fsync(fd) != 0) {

unlink_noerr("scb.tmp");

}

if (fsync(fd) != 0)

unlink_noerr("scb.tmp");

vincenzopalazzo · 2022-05-31T23:03:27Z

plugins/chanbackup.c

+	if (close(fd) != 0) {
+		unlink_noerr("scb.tmp");
+	}


Suggested change

if (close(fd) != 0) {

unlink_noerr("scb.tmp");

}

if (close(fd) != 0)

unlink_noerr("scb.tmp");

vincenzopalazzo · 2022-05-31T23:03:42Z

plugins/chanbackup.c

+	if (fd < 0) {
+		plugin_log(p, LOG_DBG, "Opening: %s", strerror(errno));
+	}


Suggested change

if (fd < 0) {

plugin_log(p, LOG_DBG, "Opening: %s", strerror(errno));

}

if (fd < 0)

plugin_log(p, LOG_DBG, "Opening: %s", strerror(errno));

vincenzopalazzo · 2022-05-31T23:05:18Z

plugins/chanbackup.c

+				    after_recover_rpc,
+				    &forward_error, NULL);
+
+	const jsmntok_t *restok = json_parse_simple(tmpctx, res, strlen(res));


we decare this on top like const jsmntok_t *restok; and after

Suggested change

const jsmntok_t *restok = json_parse_simple(tmpctx, res, strlen(res));

restok = json_parse_simple(tmpctx, res, strlen(res));

Thank you @vincenzopalazzo for the review! I've incorporated these changes in the last fixup.

rustyrussell · 2022-06-06T07:21:25Z

plugins/chanbackup.c

+#include <ccan/array_size/array_size.h>
+#include <ccan/read_write_all/read_write_all.h>
+#include <ccan/noerr/noerr.h>
+#include <ccan/err/err.h>
+#include <ccan/fdpass/fdpass.h>
+#include <ccan/tal/str/str.h>
+#include <ccan/json_out/json_out.h>
+#include <ccan/cast/cast.h>
+#include <ccan/mem/mem.h>
+#include <ccan/str/hex/hex.h>
+#include <common/json_tok.h>
+#include <common/json.h>
+#include <common/json_stream.h>
+#include <common/hsm_encryption.h>
+#include <common/type_to_string.h>


Some of these includes are probably unnecessary? (e.g. ccan/fdpass/fdpass.h)?

plugins/chanbackup.c

rustyrussell · 2022-06-06T07:27:53Z

plugins/chanbackup.c

+	if (fd < 0)
+		/* If this is not the first time we've run, it will exist. */


Always use { if you have more than one statement inside (esp. another if statement!)

plugins/chanbackup.c

rustyrussell · 2022-06-06T10:16:06Z

plugins/chanbackup.c

+	const char *scb_buf =  json_strdup(tmpctx, buf, params);
+	const jsmntok_t *scb_arr = json_parse_simple(tmpctx, scb_buf, strlen(scb_buf)),
+					*scbs = json_get_member(scb_buf, scb_arr, "scb");


You can just do const jmsntok_t *scbs = json_get_member(params, "scb"). You are converting the JSON to a string and reparsing it, which is unnecessary!

rustyrussell · 2022-06-06T10:19:46Z

plugins/chanbackup.c

+static void write_hsm(struct plugin *p, int fd, 
+					 const char *buf)
+{
+	u8 *point = tal_dup_arr(buf, u8, (u8 *)buf, strlen(buf), 0);


This routine encrypts the JSON, which is inefficient. Define a message containing an array of scbs in common/scb_wire.csv, and then use "towire_" and "fromwire_" to put it in the file.

rustyrussell · 2022-06-06T10:20:43Z

plugins/chanbackup.c

+	/* FIXME: I wanted to update the file on CHANNELD_AWAITING_LOCKIN,
+	 * But I don't get update for it, maybe because there is no previous_state,
+	 * also apparently `channel_opened` gets published when *peer* funded a channel with us? 
+	 * So, is their no way to get a notif on CHANNELD_AWAITING_LOCKIN? */


Hmm, file a bug, this seems wrong!

rustyrussell · 2022-06-06T10:21:31Z

plugins/chanbackup.c

+	 * But I don't get update for it, maybe because there is no previous_state,
+	 * also apparently `channel_opened` gets published when *peer* funded a channel with us? 
+	 * So, is their no way to get a notif on CHANNELD_AWAITING_LOCKIN? */
+	if(!strcmp(state ,"CLOSED") || !strcmp(state ,"CHANNELD_NORMAL")){


You can use json_tok_streq() BTW. And needs a space before {

rustyrussell

This is really nice! No really big feedback, just some style things and naming.

Fix those and we can enable this. I've marked it for the v0.12 milestone!

rustyrussell · 2022-07-04T03:14:34Z

lightningd/hsm_control.c

+	if (!param(cmd, buffer, params,
+		   p_req("info", param_string, &info),
+		   NULL))
+		return command_param_failed();
+
+	u8 *point = tal_dup_arr(cmd, u8, (u8*)info, strlen(info), 0);
+
+	u8 *msg = towire_hsmd_derive_secret(cmd, point);


Should we insist that info is in hex, or accept a string here? We could have two parameters (both p_opt, but make sure one is set!), called "string" and "hex": that's what datastore does.

rustyrussell · 2022-07-04T03:14:52Z

lightningd/hsm_control.c

+		   NULL))
+		return command_param_failed();
+
+	u8 *point = tal_dup_arr(cmd, u8, (u8*)info, strlen(info), 0);


point is a weird name, BTW :)

Agreed, changing it...

rustyrussell · 2022-07-04T03:15:38Z

lightningd/hsm_control.c

+	&json_getsecret,
+	"Get a pseudorandom secret key, using an info string."
+};
+AUTODATA(json_command, &getsecret_command);


You should also write doc/lightning-makesecret.7.md, and doc/schema/makesecret.request.json and doc/schema/makesecret.schema.json...

rustyrussell · 2022-07-04T03:16:13Z

tools/generate-wire.py

@@ -244,7 +244,7 @@ class Type(FieldSet):
        'tx_parts',
        'wally_psbt',
        'wally_tx',
-        'channel_type',


Don't delete channel_type here?

It's a duplicate...

rustyrussell · 2022-07-04T03:17:58Z

lightningd/channel.c

+	channel->scb->funding = *funding;
+	channel->scb->cid = *cid;
+	channel->scb->funding_sats = funding_sats;
+	channel->scb->type = (struct channel_type *) type;


Use channel_type_dup(channel->scb, type) here: it makes sense for the scb to "own" the channel_type

rustyrussell · 2022-07-04T03:31:47Z

lightningd/opening_control.c

+
+	return channel;
+}
+static struct command_result *json_commit_channel(struct command *cmd,


Rename to match the command name? i.e. json_recoverchannel.

rustyrussell · 2022-07-04T03:32:37Z

plugins/chanbackup.c

+#define VERSION ((u64)1)
+
+/* Global secret object to keep the derived encryption key for the SCB */
+struct secret secret;


rustyrussell · 2022-07-04T04:00:17Z

plugins/chanbackup.c

+	const char *scb_buf;
+	rpc_scan(p, "staticbackup",
+		 take(json_out_obj(NULL, NULL, NULL)),
+		 "{scb:%}", JSON_SCAN(json_to_str, &scb_buf));


Hmm, converting to a string an back to JSON again is really ugly :(

Can you simply call staticbackup async here? Maybe create a function to wrap this:

struct out_req *req; req = jsonrpc_request_start(plugin, NULL, "staticbackup", after_staticbackup, &forward_error, NULL); return send_outreq(plugin, req);

?

I Applied 200% of my brain and created json_to_scb_chan, which makes it easier for writing it in the file as well...

rustyrussell · 2022-07-04T04:01:22Z

plugins/chanbackup.c

+	update_scb(cmd->plugin, json_strdup(tmpctx, buf, scbs));
+	return notification_handled(cmd);


Hand this the buf and scbs token, and have it iterate through, rather than covnerting to a string and back...

This is neater, since scbs->size will tell it immediately how many scbs there are.

rustyrussell · 2022-07-04T04:02:24Z

plugins/chanbackup.c

+		"recovery",
+		"Populates the DB with stub channels",
+		"returns stub channel-id's on completion",
+		recover,


I always like the function for a command to be json_COMMANDNAME, in this case json_emergencyrecover. That way I can easily find it!

rustyrussell · 2022-07-04T20:57:23Z

Suggest adding a section to doc/BACKUP.md about this. In particular, since this now uses an atomic construct-temporary-file-and-rename, it is very safe for an automated backup process to hardlink the recovery file, copy the hardlinked name, then unlink, and the automated backup process assuredly will never acquire a corrupted copy, either.

It's also safe to just copy :) Every copy does open, read, close, which has the same semantic.

adi2011 · 2022-07-06T08:22:14Z

plugins/chanbackup.c

+	*channels = tok->size ? tal_arr(tmpctx, struct scb_chan *, tok->size) : NULL;
+
+	json_for_each_arr(i, t, tok) {
+		const u8 *scb_tmp = tal_hexdata(tmpctx, json_strdup(tmpctx, buffer, t),


Grossly violated the 80 col rule here...

niftynei · 2022-07-08T03:14:28Z

Looks like there's a bug in test_emergencybackup

2022-07-08T00:53:35.357Z **BROKEN** lightningd: FATAL SIGNAL 11 (version c4fe1d3-modded)
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: common/daemon.c:38 (send_backtrace) 0x564000429f3e
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: common/daemon.c:46 (crashdump) 0x564000429f92
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: /build/glibc-SzIz7B/glibc-2.31/signal/../sysdeps/unix/sysv/linux/x86_64/sigaction.c:0 ((null)) 0x7ff33ef5e08f
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ./bitcoin/short_channel_id.h:42 (is_stub_scid) 0x5640003a2bb0
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/channel.c:789 (channel_fail_permanent) 0x5640003a575a
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/onchain_control.c:623 (onchaind_funding_spent) 0x5640003d0496
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/peer_control.c:1638 (funding_spent) 0x5640003e5fa9
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/watch.c:271 (txowatch_fire) 0x5640004002d3
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:87 (filter_block_txs) 0x56400039fc0d
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:758 (add_tip) 0x5640003a1ac8
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:850 (get_new_block) 0x5640003a2027
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/bitcoind.c:388 (getrawblockbyheight_callback) 0x56400039e34f
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:587 (plugin_response_handle) 0x5640003f44b1
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:693 (plugin_read_json_one) 0x5640003f46da
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:738 (plugin_read_json) 0x5640003f48be
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:59 (next_plan) 0x56400049b3ce
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:407 (do_plan) 0x56400049bfd6
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:417 (io_ready) 0x56400049c018
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ccan/ccan/io/poll.c:453 (io_loop) 0x56400049e30b
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/io_loop_with_timers.c:22 (io_loop_with_timers) 0x5640003c2ce4
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: lightningd/lightningd.c:1184 (main) 0x5640003c8fc8
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: ../csu/libc-start.c:308 (__libc_start_main) 0x7ff33ef3f082
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: (null):0 ((null)) 0x56400039d38d
2022-07-08T00:53:35.357Z **BROKEN** lightningd: backtrace: (null):0 ((null)) 0xffffffffffffffff
2022-07-08T00:53:35.616Z **BROKEN** lightningd: FATAL SIGNAL 11 (version c4fe1d3-modded)
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: common/daemon.c:38 (send_backtrace) 0x562b4b73df3e
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: common/daemon.c:46 (crashdump) 0x562b4b73df92
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: /build/glibc-SzIz7B/glibc-2.31/signal/../sysdeps/unix/sysv/linux/x86_64/sigaction.c:0 ((null)) 0x7fe0516bc08f
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ./bitcoin/short_channel_id.h:42 (is_stub_scid) 0x562b4b6b6bb0
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/channel.c:789 (channel_fail_permanent) 0x562b4b6b975a
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/onchain_control.c:623 (onchaind_funding_spent) 0x562b4b6e4496
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/peer_control.c:1638 (funding_spent) 0x562b4b6f9fa9
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/watch.c:271 (txowatch_fire) 0x562b4b7142d3
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:87 (filter_block_txs) 0x562b4b6b3c0d
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:758 (add_tip) 0x562b4b6b5ac8
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/chaintopology.c:850 (get_new_block) 0x562b4b6b6027
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/bitcoind.c:388 (getrawblockbyheight_callback) 0x562b4b6b234f
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:587 (plugin_response_handle) 0x562b4b7084b1
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:693 (plugin_read_json_one) 0x562b4b7086da
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/plugin.c:738 (plugin_read_json) 0x562b4b7088be
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:59 (next_plan) 0x562b4b7af3ce
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:407 (do_plan) 0x562b4b7affd6
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ccan/ccan/io/io.c:417 (io_ready) 0x562b4b7b0018
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ccan/ccan/io/poll.c:453 (io_loop) 0x562b4b7b230b
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/io_loop_with_timers.c:22 (io_loop_with_timers) 0x562b4b6d6ce4
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: lightningd/lightningd.c:1184 (main) 0x562b4b6dcfc8
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: ../csu/libc-start.c:308 (__libc_start_main) 0x7fe05169d082
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: (null):0 ((null)) 0x562b4b6b138d
2022-07-08T00:53:35.617Z **BROKEN** lightningd: backtrace: (null):0 ((null)) 0xffffffffffffffff

niftynei · 2022-07-12T18:08:41Z

test_emergencyrecover is failing under postgres, because it expects a sqlite3 database file to delete


deftest_emergencyrecover(node_factory, bitcoind):
"""
    Test emergencyrecover
    """
        l1, l2 = node_factory.get_nodes(2, opts=[{}, {}])
        l1.rpc.connect(l2.info['id'], 'localhost', l2.port)
        c12, _ = l1.fundchannel(l2, 10**5)
        l1.stop()
>       os.unlink(os.path.join(l1.daemon.lightning_dir, TEST_NETWORK, "lightningd.sqlite3"))
E       FileNotFoundError: [Errno 2] No such file or directory: '/tmp/ltests-a1vrjh39/test_emergencyrecover_1/lightning-1/regtest/lightningd.sqlite3'

adi2011 · 2022-07-13T02:37:12Z

test_emergencyrecover is failing under postgres, because it expects a sqlite3 database file to delete

Solved...

adi2011 force-pushed the master branch 2 times, most recently from 7359161 to f84dd4b Compare May 25, 2022 07:42

cdecker reviewed May 25, 2022

View reviewed changes

rustyrussell requested changes May 25, 2022

View reviewed changes

adi2011 force-pushed the master branch from e2d724b to 379855a Compare May 27, 2022 19:27

adi2011 requested review from cdecker, rustyrussell and vincenzopalazzo May 27, 2022 20:43

adi2011 force-pushed the master branch from 6a2fec9 to 0c4cacb Compare May 27, 2022 21:03

adi2011 marked this pull request as ready for review May 28, 2022 11:30

vincenzopalazzo requested changes May 31, 2022

View reviewed changes

rustyrussell reviewed Jun 6, 2022

View reviewed changes

vincenzopalazzo mentioned this pull request Jun 12, 2022

backup seed, private key, channels #5318

Open

adi2011 force-pushed the master branch 2 times, most recently from f39b2fd to 5d2897e Compare June 17, 2022 09:23

adi2011 marked this pull request as draft June 23, 2022 22:38

adi2011 force-pushed the master branch 3 times, most recently from 184d257 to 4ed642f Compare June 24, 2022 07:05

adi2011 marked this pull request as ready for review June 24, 2022 07:07

adi2011 requested a review from niftynei June 26, 2022 06:17

niftynei added this to the v0.12 milestone Jun 27, 2022

rustyrussell requested changes Jul 4, 2022

View reviewed changes

adi2011 force-pushed the master branch from 4ed642f to 0458d2b Compare July 6, 2022 01:57

adi2011 commented Jul 6, 2022

View reviewed changes

adi2011 force-pushed the master branch 3 times, most recently from a5d88e3 to 63d85d3 Compare July 8, 2022 00:27

adi2011 force-pushed the master branch 7 times, most recently from 24ed592 to 7f33123 Compare July 8, 2022 17:07

niftynei added the needs-rebase label Jul 11, 2022

adi2011 force-pushed the master branch 3 times, most recently from ecfc697 to 6e93b98 Compare July 12, 2022 03:10

niftynei removed the needs-rebase label Jul 12, 2022

adi2011 force-pushed the master branch 2 times, most recently from 98e46fd to 77c4da9 Compare July 13, 2022 01:36

niftynei closed this Jul 13, 2022

niftynei force-pushed the master branch from 77c4da9 to f67ab2a Compare July 13, 2022 15:49

niftynei mentioned this pull request Jul 13, 2022

Static Channel Backups (we confused Github) #5422

Merged

	u8 point = tal_dup_arr(cmd, u8, (u8 )info, strlen(info), 0);
	u8 point = tal_dup_arr(cmd, u8, (u8)info, strlen(info), 0);

	const jsmntok_t *restok = json_parse_simple(tmpctx, res, strlen(res));
	restok = json_parse_simple(tmpctx, res, strlen(res));

		if (fd < 0)
		/* If this is not the first time we've run, it will exist. */

		update_scb(cmd->plugin, json_strdup(tmpctx, buf, scbs));
		return notification_handled(cmd);

Static Channel Backup. #5288

Static Channel Backup. #5288

Conversation

adi2011 commented May 24, 2022 • edited Loading

cdecker left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

adi2011 May 25, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ZmnSCPxj May 31, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

adi2011 commented May 25, 2022 • edited Loading

rustyrussell left a comment

Choose a reason for hiding this comment

adi2011 commented May 25, 2022

adi2011 commented May 27, 2022

ZmnSCPxj commented May 31, 2022

vincenzopalazzo left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rustyrussell left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rustyrussell commented Jul 4, 2022

Choose a reason for hiding this comment

niftynei commented Jul 8, 2022

niftynei commented Jul 12, 2022

adi2011 commented Jul 13, 2022 • edited Loading

adi2011 commented May 24, 2022 •

edited

Loading

adi2011 May 25, 2022 •

edited

Loading

ZmnSCPxj May 31, 2022 •

edited

Loading

adi2011 commented May 25, 2022 •

edited

Loading

adi2011 commented Jul 13, 2022 •

edited

Loading