Re: [HACKERS] raw output from copy

Andrew Dunstan wrote: > If someone can make a good case that this is going to be of > general use I'll happily go along, but I haven't seen one so far. About COPY FROM with a raw format, for instance just yesterday there was this user question on stackoverflow:

Re: [HACKERS] OOM in libpq and infinite loop with getCopyStart()

Michael Paquier writes: > On Fri, Apr 1, 2016 at 12:48 PM, Tom Lane wrote: >> Anyway, the short of my review is that we need more clarity of thought >> about what state libpq is in after a failure like this, and what that >> state looks like to the

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

Alvaro Herrera writes: > Teodor Sigaev wrote: >> may be <=>? it isn't used anywhere yet. >> >> select 'fat'::tsquery <=> 'cat'; >> select 'fat <=> cat'::tsquery; >> select 'fat <3> cat'::tsqyery; -- for non-default distance. > Dunno. That looks pretty "relationalish".

Re: [HACKERS] So, can we stop supporting Windows native now?

On Thu, Mar 31, 2016 at 09:04:35AM +0800, Craig Ringer wrote: > If we eventually get a CMake build system conversion that'll mostly go > away too. Well, maybe the good message about this new development is that autotools will start working much better on Windows and could be eventually used for

[HACKERS] syntax sugar for conditional check

Hello, I thought it would be cool to have conditional check syntax, which gets converted to simple check constraint syntax. Here’s a gist: https://gist.github.com/aj0strow/5a07f2ddcad324c4dac2c4095c821999 It’s just sugar, but i think it would make check constraints easier to read, and easier

Re: [HACKERS] OOM in libpq and infinite loop with getCopyStart()

Amit Kapila writes: > Very valid point. So, if we see with patch, I think libpq will be > in PGASYNC_COPY_XXX state after such a failure and next time when it tries > to again execute statement, it will end copy mode and then allow to proceed > from there. This design

Re: [HACKERS] SSL indicator in psql prompt

Peter Eisentraut writes: > I like how browsers show a little lock in the address bar depending on > whether SSL is in use. This could be useful in psql as well. Here is a > prototype patch. > Comments? -1 on the hard-coded UTF8, even with the encoding check (which I don't

Re: [HACKERS] dealing with extension dependencies that aren't quite 'e'

Abhijit Menon-Sen wrote: > At 2016-03-29 10:15:51 -0400, da...@pgmasters.net wrote: > > > > Either way it looks like you need to post a patch with more > > documentation - do you know when you'll have that ready? > > Here it is. > > (I was actually looking for other potential callers, but I

Re: [HACKERS] OOM in libpq and infinite loop with getCopyStart()

I wrote: > So the core of my complaint is that we need to fix things so that, whether > or not we are able to create the PGRES_FATAL_ERROR PGresult (and we'd > better consider the behavior when we cannot), ... BTW, the real Achilles' heel of any attempt to ensure sane behavior at the OOM limit is

Re: [HACKERS] snapshot too old, configured by time

Kevin Grittner wrote: > Attached is what I think you're talking about for the first patch. > AFAICS this should generate identical executable code to unpatched. > Then the patch to actually implement the feature would, instead > of adding 30-some lines with TestForOldSnapshot() would implement >

Re: [HACKERS] Logical Decoding - Execute join query

On 1 April 2016 at 17:45, Andres Freund wrote: > Hi, > > On 2016-04-01 15:09:59 +0530, hari.prasath wrote: > > I tried to execute a join query using SPI_execute() in logical > > decoding part and got inconsistent values (i am referring it as > > inconsistent

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

On Fri, Apr 1, 2016 at 5:24 PM, Tom Lane wrote: > Alvaro Herrera writes: > > Teodor Sigaev wrote: > >> may be <=>? it isn't used anywhere yet. > >> > >> select 'fat'::tsquery <=> 'cat'; > >> select 'fat <=> cat'::tsquery; > >> select 'fat <3>

[HACKERS] Re: [COMMITTERS] pgsql: Improve internationalization of messages involving type names

Moving to -hackers. Tom Lane wrote: > Alvaro Herrera writes: > > Alvaro Herrera wrote: > >> Tom Lane wrote: > >>> Actually, I think the general convention is to NOT quote type names > >>> in error messages. > > >> Ok, I'll change it. > > > Done. > > Thanks. Should

Re: [HACKERS] Performance improvement for joins where outer side is unique

David Rowley writes: > On 12 March 2016 at 11:43, Tom Lane wrote: >> It seems like the major intellectual complexity here is to figure out >> how to detect inner-side-unique at reasonable cost. I see that for >> LEFT joins you're caching that in

Re: [HACKERS] PQsendQuery+PQgetResult+PQsetSingleRowMode limitations and support

On Fri, 1 Apr 2016 05:57:33 +0200 "Shulgin, Oleksandr" wrote: > On Apr 1, 2016 02:57, "Karl O. Pinc" wrote: > > > > I assume there are no questions about supporting a > > similar functionality only without PQsetSingleRowMode, > > as follows: > >

Re: [HACKERS] syntax sugar for conditional check

Jim Nasby writes: > Rather than this, I think an exclusive-or operator would be a lot more > useful. The only difficulty I run into with CHECK constaints is when I > want to ensure that only ONE condition is true. "bool != bool" works as XOR. If you need "exactly one

Re: [HACKERS] syntax sugar for conditional check

On Thu, Mar 31, 2016 at 10:19 AM, Alexander Ostrow wrote: > Hello, > > I thought it would be cool to have conditional check syntax, which gets > converted to simple check constraint syntax. > > Here’s a gist: > > https://gist.github.com/aj0strow/5a07f2ddcad324c4dac2c4095c821999

Re: [HACKERS] syntax sugar for conditional check

On 3/31/16 12:19 PM, Alexander Ostrow wrote: Hello, I thought it would be cool to have conditional check syntax, which gets converted to simple check constraint syntax. Here’s a gist: https://gist.github.com/aj0strow/5a07f2ddcad324c4dac2c4095c821999 It’s just sugar, but i think it would make

Re: [HACKERS] syntax sugar for conditional check

2016-04-01 18:57 GMT+02:00 David G. Johnston : > On Thu, Mar 31, 2016 at 10:19 AM, Alexander Ostrow wrote: > >> Hello, >> >> I thought it would be cool to have conditional check syntax, which gets >> converted to simple check constraint syntax. >> >>

Re: [HACKERS] pgbench - remove unused clientDone parameter

Fabien wrote: > > Remove pgbench clientDone unused "ok" parameter. Seems useless, yeah, removed. > I cannot get the point of keeping a useless parameter, which is probably > there because at some point in the past it was used. If it is needed some > day it can always be reinserted. Actually it

Re: [HACKERS] More stable query plans via more predictable column statistics

"Shulgin, Oleksandr" writes: > Alright. I'm attaching the latest version of this patch split in two > parts: the first one is NULLs-related bugfix and the second is the > "improvement" part, which applies on top of the first one. I've applied the first of these

[HACKERS] pg_upgrade 9.6->9.6: column "amtype" does not exist

Hi, I'm not sure this is a bug, but before it bites back too late, I'm reporting it now. On pg_upgrading my catversion 201603111 9.6 cluster to 201603301, I got the following error: - pg_upgrade run on Fri Apr 1 22:50:07 2016

Re: [HACKERS] Fix handling of invalid sockets returned by PQsocket()

Michael Paquier wrote: > Here is v3 then, switching to "invalid socket" for those error > messages. There are two extra messages in fe-misc.c and > libpqwalreceiver.c that need a rewording that I have detected as well. Peter Eisentraut pushed this as a40814d7a. -- Álvaro Herrera

Re: [HACKERS] pg_upgrade 9.6->9.6: column "amtype" does not exist

Christoph Berg wrote: > Hi, > > I'm not sure this is a bug, but before it bites back too late, I'm > reporting it now. > > On pg_upgrading my catversion 201603111 9.6 cluster to 201603301, I > got the following error: > > - >

Re: [HACKERS] [PATCH v10] GSSAPI encryption support

Michael Paquier writes: > On Fri, Apr 1, 2016 at 12:31 PM, Robbie Harwood wrote: > >> - Fixed buffering of large replies on the serverside. This should fix >> the traceback that was being seen. The issue had to do with the >> difference

Re: [HACKERS] Speed up Clog Access by increasing CLOG buffers

On April 1, 2016 10:25:51 PM GMT+02:00, Jesper Pedersen wrote: >Hi, > >On 03/31/2016 06:21 PM, Andres Freund wrote: >> On March 31, 2016 11:13:46 PM GMT+02:00, Jesper Pedersen > wrote: >> >>> I can do a USE_CONTENT_LOCK run on 0003 if it

Re: [HACKERS] Re: [COMMITTERS] pgsql: Enable logical slots to follow timeline switches

Craig Ringer wrote: > Note that I can't use PG_GETARG_TRANSACTIONID directly since it's a macro > defined only in xid.c . It didn't seem worth extracting it and moving it to > postgres.h (where the other non-ADT-specific PG_GETARG_ macros are) or its > own new header just for this, so I've

Re: [HACKERS] pg_upgrade 9.6->9.6: column "amtype" does not exist

On Fri, Apr 1, 2016 at 2:31 PM, Christoph Berg wrote: > I'm not sure this is a bug, but before it bites back too late, I'm > reporting it now. This must be a regression from the changes made to the pg_am interface by commit 65c5fcd353a859da9e61bfb2b92a99f12937de3b. -- Peter

Re: [HACKERS] Speed up Clog Access by increasing CLOG buffers

Hi, On 03/31/2016 06:21 PM, Andres Freund wrote: On March 31, 2016 11:13:46 PM GMT+02:00, Jesper Pedersen wrote: I can do a USE_CONTENT_LOCK run on 0003 if it is something for 9.6. Yes please. I think the lock variant is realistic, the lockless did isn't. I

Re: [HACKERS] OOM in libpq and infinite loop with getCopyStart()

On Fri, Apr 1, 2016 at 10:34 AM, Michael Paquier wrote: > On Fri, Apr 1, 2016 at 12:48 PM, Tom Lane wrote: > > I thought about this patch a bit more... > > > > When I first looked at the patch, I didn't believe that it worked at > > all: it failed

Re: [HACKERS] Speedup twophase transactions

On Wed, Mar 30, 2016 at 10:19 PM, Stas Kelvich wrote: > Hm, it’s hard to create descriptive names because test changes master/slave > roles for that nodes several times during test. Really? the names used in the patch help less then. > It’s possible to call them >

Re: [HACKERS] Incorrect format in error message

On 1 April 2016 at 17:30, Tom Lane wrote: > David Rowley writes: >> The attached fixes an error message which is incorrectly using an >> unsigned format specifier instead of a signed one. > > Really though, what > astonishes me about this example

Re: [HACKERS] Move PinBuffer and UnpinBuffer to atomics

On 2016-04-01 10:35:18 +0200, Andres Freund wrote: > On 2016-04-01 13:50:10 +0530, Dilip Kumar wrote: > > I think it needs more number of runs.. After seeing this results I did not > > run head+pinunpin, > > > > Head 64 Client 128 Client > > - >

Re: [HACKERS] Timeline following for logical slots

Hi, On 2016-04-01 08:46:01 +0200, Andres Freund wrote: > That's a fundamental misunderstanding on your part (perhaps created by > imprecise docs). > > Speaking of which, did you see the proposed README I sent for > > src/backend/replication/logical ? > > I skimmed it. But given we have a CF

Re: [HACKERS] Move PinBuffer and UnpinBuffer to atomics

On Thu, Mar 31, 2016 at 5:52 PM, Andres Freund wrote: > Here's a WIP patch to evaluate. Dilip/Ashutosh, could you perhaps run > some benchmarks, to see whether this addresses the performance issues? > > I guess it'd both be interesting to compare master with master + patch, >

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

On Thu, Mar 31, 2016 at 9:14 PM, Alvaro Herrera wrote: > What led you to choose the ? operator for the FOLLOWED BY semantics? > It doesn't seem a terribly natural choice -- most other things seems to > use ? as some sort of wildcard. What about something like "...", so

Re: [HACKERS] Incorrect format in error message

On 1 April 2016 at 17:30, Tom Lane wrote: > David Rowley writes: >> The attached fixes an error message which is incorrectly using an >> unsigned format specifier instead of a signed one. > > I think that's the tip of the iceberg :-(. For

Re: [HACKERS] Incorrect format in error message

On 2016-04-01 20:18:29 +1300, David Rowley wrote: > On 1 April 2016 at 17:30, Tom Lane wrote: > > David Rowley writes: > >> The attached fixes an error message which is incorrectly using an > >> unsigned format specifier instead of a signed one.

Re: [HACKERS] IF (NOT) EXISTS in psql-completion

Hello, sorry for being a bit late. The attatched are the new version of the patch.. set. 1. 0001-Suggest-IF-NOT-EXISTS-for-tab-completion-of-psql.patch Adds IF (NOT) EXISTS completion. It doesn't fix the issue that the case of additional keywords don't follow the input. 2.

Re: [HACKERS] Move PinBuffer and UnpinBuffer to atomics

On 2016-04-01 13:50:10 +0530, Dilip Kumar wrote: > I think it needs more number of runs.. After seeing this results I did not > run head+pinunpin, > > Head 64 Client 128 Client > - > Run1 434860 356945 > Run2 275815 *275815* > Run3 437872 366560

Re: [HACKERS] Small patch: --disable-setproctitle flag

Hello, Andres > Seems more appropriate to simply manually add a #undef > HAVE_SETPROCTITLE to pg_config_manual.h in that case. Adding > configure flags for ephemeral debugger issues seems like a high churn > activity. I think you are right. OK. -- Best regards, Aleksander Alekseev

Re: [HACKERS] Logical Decoding - Execute join query

Hi, On 2016-04-01 15:09:59 +0530, hari.prasath wrote: > I tried to execute a join query using SPI_execute() in logical > decoding part and got inconsistent values (i am referring it as > inconsistent since it is returning the old values which is > present at the postgresql

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

Oleg Bartunov wrote: > On Thu, Mar 31, 2016 at 9:14 PM, Alvaro Herrera > wrote: > > > What led you to choose the ? operator for the FOLLOWED BY semantics? > > It doesn't seem a terribly natural choice -- most other things seems to > > use ? as some sort of wildcard.

Re: [HACKERS] Correction for replication slot creation error message in 9.6

On 16/04/01 8:15, Michael Paquier wrote: > On Thu, Mar 31, 2016 at 11:18 PM, Alvaro Herrera > wrote: >> Andres Freund wrote: >>> On 2016-03-31 10:15:21 +0900, Ian Barwick wrote: >> Patch changes the error message to: ERROR: replication slots can only be

[HACKERS] Logical Decoding - Execute join query

Hi all, I tried to execute a join query using SPI_execute() in logical decoding part and got inconsistent values (i am referring it as inconsistent since it is returning the old values which is present at the postgresql server start). My data directory has to tables table1(a

Re: [HACKERS] WIP: Access method extendability

Hello, Alexander > Hi! > > New revision of patches is attached. Code looks much better now, thanks. Still I believe it could be improved. I don't think that using srand() / rand() in signValue procedure the way you did is such a good idea. You create a side affect (changing current randseed)

Re: [HACKERS] dealing with extension dependencies that aren't quite 'e'

At 2016-03-29 10:15:51 -0400, da...@pgmasters.net wrote: > > Either way it looks like you need to post a patch with more > documentation - do you know when you'll have that ready? Here it is. (I was actually looking for other potential callers, but I couldn't find any. There are some places that

Re: [HACKERS] Speedup twophase transactions

On Fri, Apr 1, 2016 at 10:53 PM, Stas Kelvich wrote: > I wrote: >> While testing the patch, I found a bug in the recovery conflict code >> path. You can do the following to reproduce it: >> 1) Start a master with a standby >> 2) prepare a transaction on master >> 3) Stop

[HACKERS] Re: [COMMITTERS] pgsql: Refer to a TOKEN_USER payload as a "token user, " not as a "user

* Noah Misch (n...@leadboat.com) wrote: > On Fri, Apr 01, 2016 at 10:12:12PM -0400, Stephen Frost wrote: > > and there's no such thing as a "token user" concept. There's an enum, > > one value of which is "TokenUser" and that's what we're asking the OS to > > provide us info about, but I'd argue

[HACKERS] Re: [COMMITTERS] pgsql: Refer to a TOKEN_USER payload as a "token user, " not as a "user

* Noah Misch (n...@leadboat.com) wrote: > Refer to a TOKEN_USER payload as a "token user," not as a "user token". > > This corrects messages for can't-happen errors. The corresponding "user > token" appears in the HANDLE argument of GetTokenInformation(). I'm not at all convinced that this is

[HACKERS] Re: [COMMITTERS] pgsql: Refer to a TOKEN_USER payload as a "token user, " not as a "user

On Fri, Apr 01, 2016 at 10:12:12PM -0400, Stephen Frost wrote: > * Noah Misch (n...@leadboat.com) wrote: > > Refer to a TOKEN_USER payload as a "token user," not as a "user token". > > > > This corrects messages for can't-happen errors. The corresponding "user > > token" appears in the HANDLE

Re: [HACKERS] Re: [COMMITTERS] pgsql: Refer to a TOKEN_USER payload as a "token user, " not as a "user

Stephen Frost writes: > * Noah Misch (n...@leadboat.com) wrote: >> I see some advantages of writing "TokenUser", as you propose. However, our >> error style guide says "Avoid mentioning called function names, either; >> instead say what the code was trying to do." Mentioning

Re: [HACKERS] IF (NOT) EXISTS in psql-completion

2016-04-02 7:16 GMT+02:00 Pavel Stehule : > Hi > > 2016-04-01 10:21 GMT+02:00 Kyotaro HORIGUCHI < > horiguchi.kyot...@lab.ntt.co.jp>: > >> Hello, sorry for being a bit late. >> The attatched are the new version of the patch.. set. >> >> 1.

Re: [HACKERS] IF (NOT) EXISTS in psql-completion

Hi 2016-04-01 10:21 GMT+02:00 Kyotaro HORIGUCHI < horiguchi.kyot...@lab.ntt.co.jp>: > Hello, sorry for being a bit late. > The attatched are the new version of the patch.. set. > > 1. 0001-Suggest-IF-NOT-EXISTS-for-tab-completion-of-psql.patch > > Adds IF (NOT) EXISTS completion. It doesn't fix

Re: [HACKERS] pg_upgrade 9.6->9.6: column "amtype" does not exist

Alvaro Herrera writes: > The reason for the failure is that pg_dump knows that 9.6 installations > have the amtype column -- but on your older devel 9.6 install, it > doesn't exist. To fix it we would have to compare catalog versions in > pg_dump rather than major

Re: [HACKERS] syntax sugar for conditional check

On 4/1/16 1:08 PM, Tom Lane wrote: Jim Nasby writes: Rather than this, I think an exclusive-or operator would be a lot more useful. The only difficulty I run into with CHECK constaints is when I want to ensure that only ONE condition is true. "bool != bool" works as

Re: [HACKERS] syntax sugar for conditional check

On Fri, Apr 1, 2016 at 3:22 PM, Jim Nasby wrote: > On 4/1/16 1:08 PM, Tom Lane wrote: > >> Jim Nasby writes: >> >>> Rather than this, I think an exclusive-or operator would be a lot more >>> useful. The only difficulty I run into with CHECK

Re: [HACKERS] Using quicksort for every external sort run

On Thu, Feb 4, 2016 at 3:14 AM, Peter Geoghegan wrote: > Nyberg et al may have said it best in 1994, in the Alphasort Paper [1]: This paper is available from http://www.vldb.org/journal/VLDBJ4/P603.pdf (previously link is now dead) > The paper also has very good analysis of the

Re: [HACKERS] Support for N synchronous standby servers - take 2

On Thu, Mar 31, 2016 at 5:11 PM, Thomas Munro wrote: > On Thu, Mar 31, 2016 at 3:55 AM, Masahiko Sawada > wrote: >> On Wed, Mar 30, 2016 at 11:43 PM, Masahiko Sawada >> wrote: >>> On Tue, Mar 29, 2016 at 5:36 PM,

Re: [HACKERS] fd.c: flush data problems on osx

On Mon, Mar 21, 2016 at 03:09:56PM +0300, Stas Kelvich wrote: > On 21 Mar 2016, at 14:53, Andres Freund wrote: > > Hm. I think we should rather just skip calling pg_flush_data in the > > directory case, that very likely isn't beneficial on any OS. > > Seems reasonable,

Re: [HACKERS] [PATCH v11] GSSAPI encryption support

Hello friends, Song and dance, here's v11 both here and on my github: https://github.com/frozencemetery/postgres/tree/feature/gssencrypt11 Changes from v10: - Attempt to address a crash Michael is observing by switching to using the StringInfo/pqExpBuffer management functions over my own code

Re: [HACKERS] More stable query plans via more predictable column statistics

On Apr 1, 2016 23:14, "Tom Lane" wrote: > > "Shulgin, Oleksandr" writes: > > Alright. I'm attaching the latest version of this patch split in two > > parts: the first one is NULLs-related bugfix and the second is the > > "improvement" part,

Re: [HACKERS] pg_xlogdump fails to handle WAL file with multi-page XLP_FIRST_IS_CONTRECORD data

On 31 March 2016 at 16:13, Andres Freund wrote: > It's probably easier to just generate a humongous commit record. You can > do so by having a *lot* of subtransactions. Relatively easy to do with > plpgsql by creating them in a loop (SELECT txid_current() in EXCEPTION >

Re: [HACKERS] SSL indicator in psql prompt

On Fri, Apr 1, 2016 at 2:52 PM, Peter Eisentraut wrote: > I like how browsers show a little lock in the address bar depending on > whether SSL is in use. This could be useful in psql as well. Here is a > prototype patch. > > Example: > > Put this in .psqlrc: > > \set PROMPT1

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

På fredag 01. april 2016 kl. 15:22:55, skrev Teodor Sigaev >: > there was a character that was very similar to dots I would suggest > that.  The closest is * I think, so what do you think of "***"? And join opertator for tsqueries is the same :

Re: [HACKERS] Speedup twophase transactions

> On Apr 1, 2016, at 10:04 AM, Michael Paquier > wrote: > > I would suggest the following name modifications, node names have been > introduced to help tracking of each node's log: > - Candie => master > - Django => slave or just standby > There is no need for

Re: [HACKERS] WIP: Access method extendability

Hello > Fixed. Thanks. I don't have any other suggestions at the moment. Let see whats committers opinion on this. -- Best regards, Aleksander Alekseev http://eax.me/ -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription:

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

Well, I noticed that the docs talk about an operator that can be used in SQL (outside the tsquery parser), as well as an operator that can be Just to join 2 tsquery with operator FOLLOWED BY used inside tsquery. Inside tsquery anything would be usable, but outside that it would be good to

Re: [HACKERS] WIP: Access method extendability

Hi! On Fri, Apr 1, 2016 at 12:45 PM, Aleksander Alekseev < a.aleks...@postgrespro.ru> wrote: > Code looks much better now, thanks. Still I believe it could be improved. > > I don't think that using srand() / rand() in signValue procedure the > way you did is such a good idea. You create a side

[HACKERS] SSL indicator in psql prompt

I like how browsers show a little lock in the address bar depending on whether SSL is in use. This could be useful in psql as well. Here is a prototype patch. Example: Put this in .psqlrc: \set PROMPT1 '%s%/%R%# ' $ psql test psql (9.6devel) Type "help" for help. test=# Without SSL:

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

Teodor Sigaev wrote: > >Well, I noticed that the docs talk about an operator that can be used in > >SQL (outside the tsquery parser), as well as an operator that can be > Just to join 2 tsquery with operator FOLLOWED BY Ok. > >used inside tsquery. Inside tsquery anything would be usable, but >

Re: [HACKERS] [PATCH] Phrase search ported to 9.6

there was a character that was very similar to dots I would suggest that. The closest is * I think, so what do you think of "***"? And join opertator for tsqueries is the same : select 'fat'::tsquery *** 'cat'; ? Single '*' ? That's close to regex, any number of tokens. And it saves rules

Re: [HACKERS] Timeline following for logical slots

Hi, On 2016-04-01 13:16:18 +0800, Craig Ringer wrote: > I think it's pretty unsafe from SQL, to be sure. > > Unless failover slots get in to 9.6 we'll need to do exactly that from > internal C stuff in pglogical to support following physical failover, I know. And this makes me scared shitless.