| Subject | Re: One table become very slow |
|---|---|
| Author | marrtins_dqdp |
| Post date | 2008-09-12T21:05:13Z |
My table is very slow, not just a bit. It contains ~4000-5000 rows
(cannot tell precisly now, `couse count(*) takes tooo long)
I copied database file to linux and tried gbak - it still hangs on
CLIENTS table. But at least now I can run strace on gbak process,
which gives me a lot of (see below) (i mean A LOT - running `strace -o
output -p <pid>` gives output file 2mb in 5sec).
read(3,
"\5\00090\f\0\0\0\0\0\0\0\0\0\0\0\264P\1\0\204\0\221\0\350"..., 4096)
= 4096
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
_llseek(3, 434966528, [434966528], SEEK_SET) = 0
write(3, "\5\00090\v\0\0\0\0\0\0\0\0\0\0\0zZ\1\0\204\0\221\0\350"...,
4096) = 4096
Here starting again with read()
Has anybody any clue what this means? This output has no meaning to
me, except i see read and write to the same file - i suspect it`s
database file, not backup, because database file gets modification
time changed near to current time.
--- In firebird-support@yahoogroups.com, Svein Erling Tysvær
<svein.erling.tysvaer@...> wrote:
became slow, but in the end I created another table and filled it with
data to get performance back. It was a static table basically
containing population figures for each year, age and place - I think I
normally filled it with population data once every year.
observe something strange like this. Though my table just became a bit
slower, maybe taking one or five minutes to count the 50000 records as
opposed to subsecond or a few seconds on the other table with
identical content.
(cannot tell precisly now, `couse count(*) takes tooo long)
I copied database file to linux and tried gbak - it still hangs on
CLIENTS table. But at least now I can run strace on gbak process,
which gives me a lot of (see below) (i mean A LOT - running `strace -o
output -p <pid>` gives output file 2mb in 5sec).
read(3,
"\5\00090\f\0\0\0\0\0\0\0\0\0\0\0\264P\1\0\204\0\221\0\350"..., 4096)
= 4096
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
_llseek(3, 434966528, [434966528], SEEK_SET) = 0
write(3, "\5\00090\v\0\0\0\0\0\0\0\0\0\0\0zZ\1\0\204\0\221\0\350"...,
4096) = 4096
Here starting again with read()
Has anybody any clue what this means? This output has no meaning to
me, except i see read and write to the same file - i suspect it`s
database file, not backup, because database file gets modification
time changed near to current time.
--- In firebird-support@yahoogroups.com, Svein Erling Tysvær
<svein.erling.tysvaer@...> wrote:
>probably 1.5.3 when first observed. I have no idea how or why it
> I've experienced something similar once, but on Firebird 1.5 -
became slow, but in the end I created another table and filled it with
data to get performance back. It was a static table basically
containing population figures for each year, age and place - I think I
normally filled it with population data once every year.
>it could be nice for you to know that you're not the first person to
> I cannot help you as to how to fix problems like this, just thought
observe something strange like this. Though my table just became a bit
slower, maybe taking one or five minutes to count the 50000 records as
opposed to subsecond or a few seconds on the other table with
identical content.
>[mailto:firebird-support@yahoogroups.com] On Behalf Of marrtins_dqdp
> Set
>
> -----Original Message-----
> From: firebird-support@yahoogroups.com
> Sent: 12. september 2008 15:35
> To: firebird-support@yahoogroups.com
> Subject: [firebird-support] One table become very slow
>
> Hello!
>
> Firebird SS 2.1.1.17910-4
> Windows XP Service pack 2
> Filesize: ~800mb backup ~40mb
>
> I noticed strage behaviour with one of my databases. Suddenly one
> table (CLIENTS) become very slow to query. First it happened few weeks
> ago - basic backup/restore helped (25 Aug). Today it happened again,
> but now backup process also is going terribly slow.
>
> Have not seen/hear about such beahviour in my more than 7 year
> experience with first Interbase then Firebird.
>
> I restarted all services - apache,firebird - nothing changed.
> Restarted box - nothing changed.
>