Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
And I'm quite tempted to not flush the XLOG at all during ABORT, and to
only force synchronous commits if one of the to-be-deleted files is
non-temporary.
+1 on the first, but -1 on the second, because we'd have to track
whether
Florian G. Pflug [EMAIL PROTECTED] writes:
It seems doable, but it's not pretty. One possible scheme would be to
emit a record *after* chosing a name but *before* creating the file,
No, because the way you know the name is good is a successful
open(O_CREAT).
The idea was to log *twice*.
Florian G. Pflug [EMAIL PROTECTED] writes:
Tom Lane wrote:
+1 on the first, but -1 on the second, because we'd have to track
whether deleted files are temp or not ... it's very unclear that it'd
be worth the trouble.
I'm confused. smgrcreate already tracks if it created a file for a temp
Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
Tom Lane wrote:
+1 on the first, but -1 on the second, because we'd have to track
whether deleted files are temp or not ... it's very unclear that it'd
be worth the trouble.
I'm confused. smgrcreate already tracks if it created a
Florian G. Pflug [EMAIL PROTECTED] writes:
So should we not force synchronous commit if all to-be-deleted files
are temporary? Or is that pushing our luck too much, because it opens
the window of possible file leakage considerably wider?
I think this area is something to be left for later.
On Thu, 2007-08-30 at 00:33 -0400, Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
And I'm quite tempted to not flush the XLOG at all during ABORT, and to
only force synchronous commits if one of the to-be-deleted files is
non-temporary.
+1 on the first, but -1 on the second,
On Thu, 2007-08-30 at 14:14 -0400, Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
So should we not force synchronous commit if all to-be-deleted files
are temporary? Or is that pushing our luck too much, because it opens
the window of possible file leakage considerably wider?
Simon Riggs [EMAIL PROTECTED] writes:
-1 to second: If there are any temporary deleted files then it is
because we've done a large sort, hash join etc, so avoiding the sync
makes no difference to the overall response time.
I think you're confused, actually: this is not about temporary sort
Tom Lane wrote:
Simon Riggs [EMAIL PROTECTED] writes:
-1 to second: If there are any temporary deleted files then it is
because we've done a large sort, hash join etc, so avoiding the sync
makes no difference to the overall response time.
I think you're confused, actually: this is not
Alvaro Herrera [EMAIL PROTECTED] writes:
Why don't we create temp tables in a separate directory, anyway? That
would make it trivially easy to get rid of them during recovery: just
zap all the files in there.
Hmm ... doesn't do anything to fix the problem of file leakage for
crashes when
Tom Lane wrote:
Whether that's common enough to be worth special optimization, I'm
not sure. Your point that the transaction may have been heavyweight
enough that it'd hardly matter is still on-target, you just got there
by the wrong reasoning ;-)
My point is that there is hardly any effort
Hi
When we lazily assign XIDs, we gain another flag beside the existing
MyXactMadeTempRelUpdates, MyXactMadeXLogEntry, MyLastRecPtr and smgr's
pendingDeletes to tell what kind of actions a transaction performed. Adding
TransactionIsIsValid(GetCurrentTransactionIdIfAny()) on top of that
makes
Florian G. Pflug [EMAIL PROTECTED] writes:
I propose to do the following in my lazy XID assignment patch - can
anyone see a hole in that?
Cleaning up this area seems like a good idea. Just FYI, one reason why
there are so many LastRec pointer variables is that the WAL record
format used to
Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
I propose to do the following in my lazy XID assignment patch - can
anyone see a hole in that?
One comment is that at the time we make an entry into smgr's
pending-deletes list, I think we might not have acquired an XID yet
--- if I
Florian G. Pflug [EMAIL PROTECTED] writes:
Tom Lane wrote:
One thought here is that it's not clear that we really need a concept of
transaction-controlled vs not-transaction-controlled xlog records
anymore.
I've thinking about keeping XLOG_NO_TRAN, and doing
if (!no_tran)
Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
Tom Lane wrote:
One thought here is that it's not clear that we really need a concept of
transaction-controlled vs not-transaction-controlled xlog records
anymore.
I've thinking about keeping XLOG_NO_TRAN, and doing
if (!no_tran)
Tom Lane wrote:
One comment is that at the time we make an entry into smgr's
pending-deletes list, I think we might not have acquired an XID yet
--- if I understand your patch correctly, a CREATE TABLE would acquire
an XID when it makes its first catalog insertion, and that happens
after
Florian G. Pflug [EMAIL PROTECTED] writes:
One comment is that at the time we make an entry into smgr's
pending-deletes list, I think we might not have acquired an XID yet
Hm.. I was just going to implement this, but I'm now wondering if
thats really worth it.
Basically what you'd give up is
Tom Lane wrote:
Florian G. Pflug [EMAIL PROTECTED] writes:
One comment is that at the time we make an entry into smgr's
pending-deletes list, I think we might not have acquired an XID yet
Hm.. I was just going to implement this, but I'm now wondering if
thats really worth it.
Basically
Florian G. Pflug [EMAIL PROTECTED] writes:
And I'm quite tempted to not flush the XLOG at all during ABORT, and to
only force synchronous commits if one of the to-be-deleted files is
non-temporary.
+1 on the first, but -1 on the second, because we'd have to track
whether deleted files are temp
20 matches
Mail list logo