I've run into a problem with a PITR setup at a client. The problem is that
whenever the CIFS NAS device that we're mounting at /mnt/pgbackup has
problems, it seems that the current client connection gets blocked and this
eventually builds up to a "sorry, too many clients already" error. I'm
w
What might be more bullet proof would be to make the archive command
copy the file to an intermediate local directory, then have a
daemon/cron job that wakes up once a minute or so, check for new files,
then copy them to the network mount. You may want to use something like
lofs to make sure t
Jeff Frost <[EMAIL PROTECTED]> writes:
> I've run into a problem with a PITR setup at a client. The problem is that
> whenever the CIFS NAS device that we're mounting at /mnt/pgbackup has
> problems, it seems that the current client connection gets blocked and this
> eventually builds up to a "
On Mon, 15 May 2006, Tom Lane wrote:
No, I can't see what the connection should be there. It's supposed to
be designed so that the archive command can take its sweet old time and
nothing happens except that a backlog of WAL files builds up in pg_xlog.
That's what I thought, but that doesn't s
Jeff Frost <[EMAIL PROTECTED]> writes:
> On Mon, 15 May 2006, Tom Lane wrote:
>> No, I can't see what the connection should be there. It's supposed to
>> be designed so that the archive command can take its sweet old time and
>> nothing happens except that a backlog of WAL files builds up in pg_xl
On Mon, 15 May 2006, Tom Lane wrote:
That's what I thought, but that doesn't seem to be what I'm observing. Of
course the NAS device only gets wedged about once every month or two, so it's
difficult to reproduce.
If it's really a PG bug, it should be trivial to reproduce: put a long
sleep in
On Mon, 2006-05-15 at 09:28 -0700, Jeff Frost wrote:
> I've run into a problem with a PITR setup at a client. The problem is that
> whenever the CIFS NAS device that we're mounting at /mnt/pgbackup has
> problems
What kind of problems?
> , it seems that the current client connection gets block
On Mon, 15 May 2006, Simon Riggs wrote:
On Mon, 2006-05-15 at 09:28 -0700, Jeff Frost wrote:
I've run into a problem with a PITR setup at a client. The problem is that
whenever the CIFS NAS device that we're mounting at /mnt/pgbackup has
problems
What kind of problems?
It becomes unwritabl
On Mon, 2006-05-15 at 16:29, Jeff Frost wrote:
> On Mon, 15 May 2006, Simon Riggs wrote:
>
> > On Mon, 2006-05-15 at 09:28 -0700, Jeff Frost wrote:
> >> I've run into a problem with a PITR setup at a client. The problem is that
> >> whenever the CIFS NAS device that we're mounting at /mnt/pgbacku
On Mon, 2006-05-15 at 14:29 -0700, Jeff Frost wrote:
> On Mon, 15 May 2006, Simon Riggs wrote:
>
> > On Mon, 2006-05-15 at 09:28 -0700, Jeff Frost wrote:
> >> I've run into a problem with a PITR setup at a client. The problem is that
> >> whenever the CIFS NAS device that we're mounting at /mnt/p
On Mon, 2006-05-15 at 16:58 -0700, Jeff Frost wrote:
> The log is below. Note that the problem began around 2a.m. around the time
> the complaint about checkpoint segments happens. After a bit of research it
> appears that the checkpoint complaint happens when our db maintenance job
> kicks o
On Tue, 16 May 2006, Simon Riggs wrote:
I don't see much evidence for a connection between archiver and these
issues. The problems start after autovacuum of "vb_web" at 02:08. That
seems much more likely to have something to do with client connections
than the archiver - which is really nothing
Jeff Frost <[EMAIL PROTECTED]> writes:
> On Tue, 16 May 2006, Simon Riggs wrote:
>> Whatever happened between 02:08 and 02:14 seems important.
> I have the logs and after reviewing /var/log/messages for that time period,
> there is no other activity besides postgres.
I have a lurking feeling tha
On Wed, 17 May 2006, Tom Lane wrote:
I have a lurking feeling that the still-hypothetical connection between
archiver and foreground operations might come into operation at pg_clog
page boundaries (which require emitting XLOG events) --- that is, every
32K transactions something special happens
On Wed, 2006-05-17 at 00:36 -0400, Tom Lane wrote:
> Jeff Frost <[EMAIL PROTECTED]> writes:
> > On Tue, 16 May 2006, Simon Riggs wrote:
> >> Whatever happened between 02:08 and 02:14 seems important.
>
> > I have the logs and after reviewing /var/log/messages for that time period,
> > there is no
Simon Riggs <[EMAIL PROTECTED]> writes:
> You'll have to explain a little more. I checked the archives...
I was thinking of
http://archives.postgresql.org/pgsql-hackers/2004-01/msg00530.php
full explanation here:
http://archives.postgresql.org/pgsql-hackers/2004-01/msg00606.php
> The "lurking fee
On Wed, 2006-05-17 at 10:01 -0400, Tom Lane wrote:
> Simon Riggs <[EMAIL PROTECTED]> writes:
> > You'll have to explain a little more. I checked the archives...
>
> I was thinking of
> http://archives.postgresql.org/pgsql-hackers/2004-01/msg00530.php
> full explanation here:
> http://archives.post
On Wed, 17 May 2006, Tom Lane wrote:
The "lurking feeling" scenario above might or might nor be an issue
here, but I can't see how the archiver could be involved at all.
Well, I don't see it either; at this point we're waiting on Jeff to
provide some harder evidence ...
Was the 3,000 transac
Jeff Frost <[EMAIL PROTECTED]> writes:
> Was the 3,000 transactions per minute helpful? What other evidence should I
> be looking for?
Did you try generating a test case using a long sleep() as a replacement
for the archive_command script? If there is a PG bug here it shouldn't
be that hard to
On Wed, 17 May 2006, Tom Lane wrote:
Did you try generating a test case using a long sleep() as a replacement
for the archive_command script? If there is a PG bug here it shouldn't
be that hard to expose it in a simple test case. I'm up to my armpits
in other stuff and don't have time to try i
On Wed, 17 May 2006, Jeff Frost wrote:
On Wed, 17 May 2006, Tom Lane wrote:
Did you try generating a test case using a long sleep() as a replacement
for the archive_command script? If there is a PG bug here it shouldn't
be that hard to expose it in a simple test case. I'm up to my armpits
in
Jeff Frost <[EMAIL PROTECTED]> writes:
> I seem to get alot of these:
> May 17 21:34:04 discord postgres[20573]: [5-1] WARNING: could not rename
> file
> "pg_xlog/archive_status/00010001.ready" to
> May 17 21:34:04 discord postgres[20573]: [5-2]
> "pg_xlog/archive_status/00
On Thu, 18 May 2006, Tom Lane wrote:
Jeff Frost <[EMAIL PROTECTED]> writes:
I seem to get alot of these:
May 17 21:34:04 discord postgres[20573]: [5-1] WARNING: could not rename file
"pg_xlog/archive_status/00010001.ready" to
May 17 21:34:04 discord postgres[20573]: [5-2]
"p
On Wed, 17 May 2006, Jeff Frost wrote:
And in the window where I started postgres via pg_ctl, I had this:
cat: pg_xlog/0001000E: No such file or directory
cat: pg_xlog/0001000E: No such file or directory
Hrmmm...my pgbench died with an integer out of range erro
On Wed, 2006-05-17 at 22:45 -0700, Jeff Frost wrote:
> On Thu, 18 May 2006, Tom Lane wrote:
>
> > Jeff Frost <[EMAIL PROTECTED]> writes:
> >> I seem to get alot of these:
> >
> >> May 17 21:34:04 discord postgres[20573]: [5-1] WARNING: could not rename
> >> file
> >> "pg_xlog/archive_status/
Jeff Frost <[EMAIL PROTECTED]> writes:
> Hrmmm...my pgbench died with an integer out of range error:
That's normal, if you run it long enough without re-creating the tables.
It keeps adding small values to the balances, and eventually they
overflow. (Possibly someone should fix it so that the del
On Thu, 18 May 2006, Simon Riggs wrote:
Seems so.
Can you post the full test, plus full execution log.
[You don't need to "cat" you could just do "ls" instead FWIW]
Are you doing *anything* with pg_xlog directory or below? I understand
your saying No to that question and pg_xlog has not been
On Thu, 2006-05-18 at 10:08 -0700, Jeff Frost wrote:
> May 18 08:00:18 discord postgres[20228]: [129-1] LOG: archived transaction
> log file "0001007F"
> May 18 08:00:41 discord postgres[20573]: [254-1] LOG: archived transaction
> log file "0001007F"
> May 18 08
On Fri, 19 May 2006, Tom Lane wrote:
Jeff Frost <[EMAIL PROTECTED]> writes:
Do you think the postmaster on 5432 is trying to archive the other
postmaster's WAL files somehow?
Not as long as they aren't in the same data directory ;-). What Simon
was wondering about was whether an archiver pro
On Fri, 19 May 2006, Tom Lane wrote:
Well, there's our smoking gun. IIRC, all the failures you showed us are
consistent with race conditions caused by multiple archiver processes
all trying to do the same tasks concurrently.
Do you frequently stop and restart the postmaster? Because I don't s
Jeff Frost <[EMAIL PROTECTED]> writes:
> Hurray! Unfortunately, the postmaster on the original troubled server almost
> never gets restarted, and in fact only has only one archiver process running
> right now. Drat!
Well, the fact that there's only one archiver *now* doesn't mean there
wasn't m
I wrote:
> Well, the fact that there's only one archiver *now* doesn't mean there
> wasn't more than one when the problem happened. The orphaned archiver
> would eventually quit.
But, actually, nevermind: we have explained the failures you were seeing
in the test setup, but a multiple-active-arch
On Fri, 2006-05-19 at 12:20 -0400, Tom Lane wrote:
> I wrote:
> > Well, the fact that there's only one archiver *now* doesn't mean there
> > wasn't more than one when the problem happened. The orphaned archiver
> > would eventually quit.
>
> But, actually, nevermind: we have explained the failure
On Fri, 19 May 2006, Tom Lane wrote:
Well, the fact that there's only one archiver *now* doesn't mean there
wasn't more than one when the problem happened. The orphaned archiver
would eventually quit.
Do you have logs that would let you check when the production postmaster
was restarted?
I l
On Fri, 19 May 2006, Tom Lane wrote:
What I'd suggest is resuming the test after making sure you've killed
off any old archivers, and seeing if you can make any progress on
reproducing the original problem. We definitely need a
multiple-archiver interlock, but I think that must be unrelated to
On Fri, 2006-05-19 at 09:36 -0700, Jeff Frost wrote:
> On Fri, 19 May 2006, Tom Lane wrote:
>
> > What I'd suggest is resuming the test after making sure you've killed
> > off any old archivers, and seeing if you can make any progress on
> > reproducing the original problem. We definitely need a
On Fri, 19 May 2006, Simon Riggs wrote:
Now I can run my same pg_bench, or do you guys
have any other suggestions on attempting to reproduce the problem?
No. We're back on track to try to reproduce the original error.
I've been futzing with trying to reproduce the original problem for a few
On Sun, 2006-05-21 at 14:16 -0700, Jeff Frost wrote:
> On Fri, 19 May 2006, Simon Riggs wrote:
>
> >> Now I can run my same pg_bench, or do you guys
> >> have any other suggestions on attempting to reproduce the problem?
> >
> > No. We're back on track to try to reproduce the original error.
>
>
On Sun, 21 May 2006, Simon Riggs wrote:
I've been futzing with trying to reproduce the original problem for a few days
and so far postgres seems to be just fine with a long delay on archiving, so
now I'm rather at a loss. In fact, I currently have 1,234 xlog files in
pg_xlog, but the archiver i
On Sun, 21 May 2006, Jeff Frost wrote:
So the chances of the original problem being archiver related are
receding...
This is possible, but I guess I should try and reproduce the actual problem
with the same archive_command script and a CIFS mount just to see what
happens. Perhaps the real r
Jeff Frost <[EMAIL PROTECTED]> writes:
> I tried both pulling the plug on the CIFS server and unsharing the CIFS
> share,
> but pgbench continued completely unconcerned. I guess the failure mode of
> the
> NAS device in the customer colo must be something different that I don't yet
> know how
On Tue, 23 May 2006, Tom Lane wrote:
I'm still thinking that the simplest explanation is that $PGDATA/pg_clog/
is on the NAS device. Please double-check the file locations.
I know that seems like an excellent candidate, but it really isn't, I swear.
In fact, you almost had me convinced the l
On Fri, 2006-05-19 at 08:53 -0700, Jeff Frost wrote:
> On Fri, 19 May 2006, Tom Lane wrote:
>
> > Jeff Frost <[EMAIL PROTECTED]> writes:
> >> Do you think the postmaster on 5432 is trying to archive the other
> >> postmaster's WAL files somehow?
> >
> > Not as long as they aren't in the same data
Jeff Frost <[EMAIL PROTECTED]> writes:
> Do you think the postmaster on 5432 is trying to archive the other
> postmaster's WAL files somehow?
Not as long as they aren't in the same data directory ;-). What Simon
was wondering about was whether an archiver process had somehow been
left over from
On Fri, 2006-05-19 at 08:23 -0700, Jeff Frost wrote:
> On Fri, 19 May 2006, Simon Riggs wrote:
>
> > On Thu, 2006-05-18 at 10:08 -0700, Jeff Frost wrote:
> >
> >> May 18 08:00:18 discord postgres[20228]: [129-1] LOG: archived
> >> transaction log file "0001007F"
> >> May 18 08:00
On Fri, 19 May 2006, Simon Riggs wrote:
On Thu, 2006-05-18 at 10:08 -0700, Jeff Frost wrote:
May 18 08:00:18 discord postgres[20228]: [129-1] LOG: archived transaction log file
"0001007F"
May 18 08:00:41 discord postgres[20573]: [254-1] LOG: archived transaction log file
"0
Jeff Frost <[EMAIL PROTECTED]> writes:
> Well now, will you look at this:
> postgres 20228 1 0 May17 ?00:00:00 postgres: archiver process
> postgres 20573 1 0 May17 ?00:00:00 postgres: archiver process
> postgres 23817 23810 0 May17 pts/11 00:00:00 postgres: archiver p
47 matches
Mail list logo