Обсуждение: 7.3.2 strange behavior (long)

Поиск
Список
Период
Сортировка

7.3.2 strange behavior (long)

От
andrea suisani
Дата:
Hello,

Firstly I apologise for my terrible english.
we are using postgresql 7.2.x (currently 7.2.4)
on our production server for last year (and also
  we are on suse 7.2).

Last night i was working on the project for
porting our application on pgsql 7.3.2 when a strange
thing happen. The machine hangs (not only postgres) after
I submit via psql this query "update coop_uni set msg_quote=''"
to the server. The structure of this table is:

coop2003=> \d coop_uni
                  Table "public.coop_uni"
     Column    |           Type           |   Modifiers
--------------+--------------------------+---------------
  prefisso     | character varying(6)     |
  telefono     | character varying(20)    |
  cap          | character varying(6)     |
  cod_com      | integer                  |
  cod_prov     | integer                  |
  citta        | character(35)            |
  nome_cognome | character varying(70)    |
  capoluo      | integer                  |
  estrazione   | timestamp with time zone |
  id           | integer                  |
  rand         | double precision         |
  num_estra    | integer                  | default 0
  occupato     | boolean                  | default false
  tipo         | integer                  |
  strato       | integer                  |
  sample       | integer                  |
  msg_quote    | character varying(20)    |
Indexes: id_tipo_uni_idx btree (id, tipo),
          iduni_idx btree (id),
          pre_tel_idx btree (prefisso, telefono),
          prov_capo_rand_idx btree (cod_prov, capoluo, rand),
          strato_idx btree (tipo, strato, sample)


follow the log of postgres during crash:

[snip]

2003-03-19 23:48:20 [3840]   LOG:  query: select oid,* from coop_uni ;
2003-03-19 23:48:20 [3840]   DEBUG:  ProcessQuery
2003-03-19 23:48:23 [3840]   DEBUG:  CommitTransactionCommand
2003-03-19 23:48:23 [3840]   LOG:  duration: 3.048701 sec
2003-03-19 23:48:35 [3840]   DEBUG:  StartTransactionCommand
2003-03-19 23:48:35 [3840]   LOG:  query: select oid,* from coop_uni order by oid;
2003-03-19 23:48:35 [3840]   DEBUG:  ProcessQuery
2003-03-19 23:48:39 [3840]   DEBUG:  CommitTransactionCommand
2003-03-19 23:48:39 [3840]   LOG:  duration: 4.403131 sec
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@þþ
þþþþþþþþþþeggggggÿoýefÿÿÿÿ^@^@^@^@^@^X~_@ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿþþþþþþþÿþþþþþþþÿÿfÿÿiÿÿÿÿýýfýýff
^@^@^@^@^@Ø~Q@^@^@^@^@^@^T~T@^@^@^@^@À~Yá@^@^@^@^@^@Óý@þþþþfýýþ   3111957
ÿþþÿþþiý^@^@^@^@^@Àj@ýýýýýþþeist professionale per il commercio      ef~G~G~Ggÿýimresa
aýýþþþþýýrigianale       ^@^@^@^@^@| ~C@^@^@^@^@^@^W°@ýýþþþþþ÷amministrazione
fÿeÿþþþþjþþþþþþþþþþþþþeggggggÿeý^@^@^@^@^@8~_@hfÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿþþþþþþþÿþþþþþþþÿÿfÿÿiÿÿÿÿý^@^@^@^@^@|
~C@÷jýýffþþ^@^@^@^@À~Tá@^@^@^@^@~@Qç@þþfýýþÿþ   3071962 þÿþþhýýý^@^@^@^@^@^@l@ragioneria
     þþþþþeefli~Lgÿýþþufficio þþþþýýýý^@^@^@^@^@\~R@^@^@^@^@^@^W°@amministrazione
þþþþþ÷fÿeÿþþþþeþþþþþþþþþþþþþeggggggÿiýhf^@^@^@^@^@,~_@ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿþþþþþþþÿþþþþþþþÿÿ
fÿÿiÿÿÿÿý÷g^@^@^@^@^@\~R@ýýeeþþþþ^@^@^@^@^@~Yá@^@^@^@^@~@^@×@eýýþÿþþÿ


When i try to manually reboot the box
I have another problem. the machine hangs during
the boot with this message:

"Unable to handle kernel paging request at virtual address dc7d00
printing eip:
c01ad272
* pde = 17edb...
  ....
  Oops:  000
  CPU: 0

  ...."

I've tried again and after the third manual reboot
the machine boot correctly. After that I start postgres7.3.2.
This is the log:

DEBUG:  /opt/postgres-7.3.2/bin/postmaster: PostmasterMain: initial environ dump:
2003-03-19 23:56:09 [601]    DEBUG:  -----------------------------------------
2003-03-19 23:56:09 [601]    DEBUG:     PWD=/var/lib/pgsql
2003-03-19 23:56:09 [601]    DEBUG:     PAGER=less
2003-03-19 23:56:09 [601]    DEBUG:     HOSTNAME=suino
2003-03-19 23:56:09 [601]    DEBUG:     LESSCLOSE=lessclose.sh %s %s
2003-03-19 23:56:09 [601]    DEBUG:     LS_OPTIONS=-N --color=tty -T 0

[snip]

2003-03-19 23:56:09 [601]    DEBUG:     LC_COLLATE=POSIX
2003-03-19 23:56:09 [601]    DEBUG:     _=/opt/postgres-7.3.2/bin/postmaster
2003-03-19 23:56:09 [601]    DEBUG:  -----------------------------------------
2003-03-19 23:56:09 [601]    DEBUG:  FindExec: found "/opt/postgres-7.3.2/bin/postgres"
using argv[0]
2003-03-19 23:56:09 [601]    DEBUG:  invoking IpcMemoryCreate(size=1466368)
2003-03-19 23:56:09 [601]    DEBUG:  FindExec: found "/opt/postgres-7.3.2/bin/postmaster"
using argv[0]
2003-03-19 23:56:09 [605]    LOG:  database system was interrupted at 2003-03-19 23:49:59 CET
2003-03-19 23:56:09 [605]    LOG:  checkpoint record is at 0/7FDABC0
2003-03-19 23:56:09 [605]    LOG:  redo record is at 0/7FDABC0; undo record is at 0/0;
shutdown FALSE
2003-03-19 23:56:09 [605]    LOG:  next transaction id: 1446; next oid: 222355
2003-03-19 23:56:09 [605]    LOG:  database system was not properly shut down; automatic
recovery in progress
2003-03-19 23:56:09 [605]    LOG:  redo starts at 0/7FDAC00
2003-03-19 23:56:20 [605]    LOG:  ReadRecord: record with zero length at 0/93508AC
2003-03-19 23:56:20 [605]    LOG:  redo done at 0/9350870
2003-03-19 23:56:24 [605]    LOG:  recycled transaction log file 0000000000000006
2003-03-19 23:56:24 [605]    LOG:  database system is ready


Anyways the db design is very simple
(no trigger, no fks, and there are only few
store procs written in plpgsql)


The developing box is a "small" machine
(celeron 1.2GHZ,ram 384MB, no scsi disks and
PGDATA is on one ide disk), the kernel version
is 2.4.20

I wanted to ask to you if this problem can have had
to a hardware issue or to something else.


Andrea Suisani


other info:

- sickpig@suino:postgres7.3.2 $ pg_config --configure
'--prefix=/opt/postgres-7.3.2/' '--with-perl' '--with-tcl' '--with-tk' '--enable-syslog'

- the file $PGDATA/postgresql.conf is the one that come from intallation without any
modification

- cat $PGDATA/postmaster.opts:
/opt/postgres-7.3.2/bin/postmaster '-p' '5433' '-d' '2' '-i'



Re: 7.3.2 strange behavior (long)

От
Robert Treat
Дата:
On Thu, 2003-03-20 at 08:05, andrea suisani wrote:
> Hello,
>
> Firstly I apologise for my terrible english.
> we are using postgresql 7.2.x (currently 7.2.4)
> on our production server for last year (and also
>   we are on suse 7.2).
>
> Last night i was working on the project for
> porting our application on pgsql 7.3.2 when a strange
> thing happen. The machine hangs (not only postgres) after
> I submit via psql this query "update coop_uni set msg_quote=''"
> to the server. The structure of this table is:
>
<snip>
>
> When i try to manually reboot the box
> I have another problem. the machine hangs during
> the boot with this message:
>
> "Unable to handle kernel paging request at virtual address dc7d00
> printing eip:
> c01ad272
> * pde = 17edb...
>   ....
>   Oops:  000
>   CPU: 0
>
>   ...."
>
<snip>
>
> The developing box is a "small" machine
> (celeron 1.2GHZ,ram 384MB, no scsi disks and
> PGDATA is on one ide disk), the kernel version
> is 2.4.20
>
> I wanted to ask to you if this problem can have had
> to a hardware issue or to something else.
>

I'm almost certain you have some kind of hardware issue, though
unfortunately I can't deduce just what it is. I'd start with checking
your ram, then maybe hdd. HTH.

Robert Treat


Re: 7.3.2 strange behavior (long)

От
andrea suisani
Дата:
thanks for the reply.
I've checked the RAM with memtest
and nothing seems to be broken  with it.

Also hd seems to be ok.

Anyways in the last few days I've heavily
stressed the box with pgsql ver 7.2.4
and nothing went wrong.

ASAP I will try again with pgsql ver 7.3.2


thanks a lot

Andrea Suisani

Robert Treat wrote:
> On Thu, 2003-03-20 at 08:05, andrea suisani wrote:
>

<snip>

>
>
> I'm almost certain you have some kind of hardware issue, though
> unfortunately I can't deduce just what it is. I'd start with checking
> your ram, then maybe hdd. HTH.
>
> Robert Treat
>