Would you be willing to try to load a large table (from a SQL file). It reliably demonstrates the behavior causing an "invalid page header" during completion (for me)
It seems (to me) that while WAL logs are being flushed, a certain type of failed SQL insert will cause a page to be marked wrong.
I have not been able to nail down the problem but confirmation that another user can replicate the problem with the exact dataset would greatly improve the ability to get this fixed.
BTW I too am using XFS so that may also be the culprit. any non XFS users care to try?
C.
Hi list,>
I am working with 7.4.1 under Linux (SuSE 8.1) The server is a HP > ProLiant> DL 380-G3, 2x Intel Pentium4-Xeon, 2.8 GHz, 4 GB memory and a RAID 5 > system> with ca. 500 GB diskspace (xfs file system)>
When doing big transactions or changes (UPDATE several million rows> in one> step) on a database with ca. 50 GB diskspace and ca 15 million > entries,> PostGIS and lots of indexes, I get errors like>
ERROR: invalid page header in block 582024 of relation ...>
The error does not occur regularly. It seems to me that the error is > related> with heavy load and heavy I/O . CPU and memory does not seem the > problem,> according to my hotsanic tools.>
which suggest to me that I have run into hardware trouble (disk? > memory?!)> But I supposed that my server was not SO bad for this database> application...>
How to fix that? (I mean - not to repair the tables or pages but to > get a> stable one? ) Trying another sotrage hardware?!> --> -------------------------------------------------------------------------------> Dr.-Ing. Hubert Fröhlich> Bezirksfinanzdirektion München> Alexandrastr. 3, D-80538 München, GERMANY> Tel. :+49 (0)89 / 2190 - 2980> Fax :+49 (0)89 / 2190 - 2997> hubert dot froehlich at bvv dot bayern dot de>
---------------------------(end of broadcast)--------------------------- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
Hubert Fröhlich 18 June 2004 12:29:12 [ permanent link ]
Carl,>
Hubert,>
I too have noticed such an issue>
Would you be willing to try to load a large table (from a SQL file).> It reliably demonstrates the behavior "invalid page header"> after completion (for me)
Up to now, my phenomenon is a) described as in my posting below The error occurs irregularly, mostly when UPDATinig large (10 mill. entries) tables, especially when doing several million UPDATES in ONE transaction. After that, I reduced the transaction size to < 300000 updates per transaction, and the error still occurs. b) Meanwhile I have "managed" to reproduce the error on a smaller scale rather reliably. (somewhat similar to your scenario) First, I INSERTed ca. 3 mill. rows into a database in transaction packets of 100 each. So far, all fine. But when doing the same job simultaneously on two databases in the same cluster on the same server, I got the error. c) I conclude: The issue is not an issue of too big transactions. However, might it be an issue of coordinating several databases on the same cluster?
It seems to me that while there are WAL logs being flushed, a certain > type of failed SQL insert will cause a page to be marked wrong.>
I have not been able to nail down the problem but confirmation that > another user can replicate the problem with the exact dataset would > greatly improve the ability to get this fixed. >
I'd like to test my server(s) with your data, so send'em.
BTW I too am using XFS so that may also be the culprit. any non XFS > users care to try?
I want to test with a different storage device but i haven't one yet.
Greetings,
Hubert
C.>
Hi list,>>
I am working with 7.4.1 under Linux (SuSE 8.1) The server is a HP >> ProLiant DL 380-G3, 2x Intel Pentium4-Xeon, 2.8 GHz, 4 GB memory and a >> RAID 5 system with ca. 500 GB diskspace (xfs file system)>>
When doing big transactions or changes (UPDATE several million rows in >> one step) on a database with ca. 50 GB diskspace and ca 15 million >> entries, PostGIS and lots of indexes, I get errors like>>
ERROR: invalid page header in block 582024 of relation ...>>
The error does not occur regularly. It seems to me that the error is >> related with heavy load and heavy I/O . CPU and memory does not seem >> the problem, according to my hotsanic tools.>>
which suggest to me that I have run into hardware trouble (disk? >> memory?!) But I supposed that my server was not SO bad for this >> database application...>>
How to fix that? (I mean - not to repair the tables or pages but to >> get a stable one? ) Trying another sotrage hardware?!
Carl Anderson <carl.anderson@vadose.org> writes:> Would you be willing to try to load a large table (from a SQL file).> It reliably demonstrates the behavior causing an "invalid page header"> during completion (for me)
I would very much like to see a test case that reliably produces "invalid page header" ...
regards, tom lane
---------------------------(end of broadcast)--------------------------- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
=?ISO-8859-1?Q?Hubert_Fr=F6hlich?= <hubert.froehlich@bvv.bayern.de> writes:> E.g. On a client (Athlon 1400+, 512 RAM, XFS, SuSE Linux 8.1 with 2.4.19 > kernel), I regularly crash when using> numberOfRows=300000 and ndatabases=3
FWIW, I could not reproduce any such problem here. I ran three test cycles with numberOfRows=300000 and ndatabases=3 under PG 7.4.3 and another three under CVS tip (amounting to something over six hours total runtime). I saw no errors whatsoever. I was using Red Hat 8.0 with 2.4.18-24.8.0 kernel on a Dell P4 desktop machine.
I'm leaning to the theory that you've got intermittent hardware problems.
regards, tom lane
---------------------------(end of broadcast)--------------------------- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to majordomo@postgresql.org so that your message can get through to the mailing list cleanly
Manfred Koizar 20 August 2004 18:23:12 [ permanent link ]
On Fri, 23 Jul 2004 14:33:50 -0400, Tom Lane <tgl@sss.pgh.pa.us> wrote:>=?ISO-8859-1?Q?Hubert_Fr=F6hlich?= <hubert.froehlich@bvv.bayern.de> writes:>> E.g. On a client (Athlon 1400+, 512 RAM, XFS, SuSE Linux 8.1 with 2.4.19 >> kernel), I regularly crash [...]
Hubert, have you ever tested with something different than XFS? I volunteer to cross-check any test you want me to, as long as you can reproduce the problem on an ext2 file system first.
FWIW, I could not reproduce any such problem here. [...]> I was using Red Hat 8.0>with 2.4.18-24.8.0 kernel on a Dell P4 desktop machine.
Tom, do you remember what type of FS you used?
Servus Manfred
---------------------------(end of broadcast)--------------------------- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
Manfred Koizar <mkoi-pg@aon.at> writes:> On Fri, 23 Jul 2004 14:33:50 -0400, Tom Lane <tgl@sss.pgh.pa.us> wrote:>> FWIW, I could not reproduce any such problem here. [...]>> I was using Red Hat 8.0>> with 2.4.18-24.8.0 kernel on a Dell P4 desktop machine.
Tom, do you remember what type of FS you used?
That machine is using ext3.
regards, tom lane
---------------------------(end of broadcast)--------------------------- TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
If you would like to report an abuse of our service, such as a spam message, please . Если Вы хотите пожаловаться на содержимое этой страницы, пожалуйста .
