Received: from maia.hub.org (maia-3.hub.org [200.46.204.243]) by mail.postgresql.org (Postfix) with ESMTP id 2F3D51337B90 for ; Mon, 1 Nov 2010 13:10:21 -0300 (ADT) Received: from mail.postgresql.org ([200.46.204.86]) by maia.hub.org (mx1.hub.org [200.46.204.243]) (amavisd-maia, port 10024) with ESMTP id 92280-02 for ; Mon, 1 Nov 2010 16:10:14 +0000 (UTC) X-Greylist: domain auto-whitelisted by SQLgrey-1.7.6 Received: from nm30-vm0.bullet.mail.ne1.yahoo.com (nm30-vm0.bullet.mail.ne1.yahoo.com [98.138.91.69]) by mail.postgresql.org (Postfix) with SMTP id B5D3A1337B8F for ; Mon, 1 Nov 2010 13:10:12 -0300 (ADT) Received: from [98.138.90.55] by nm30.bullet.mail.ne1.yahoo.com with NNFMP; 01 Nov 2010 16:10:11 -0000 Received: from [98.138.87.11] by tm8.bullet.mail.ne1.yahoo.com with NNFMP; 01 Nov 2010 16:10:10 -0000 Received: from [127.0.0.1] by omp1011.mail.ne1.yahoo.com with NNFMP; 01 Nov 2010 16:08:18 -0000 X-Yahoo-Newman-Property: ymail-3 X-Yahoo-Newman-Id: 255856.8537.bm@omp1011.mail.ne1.yahoo.com Received: (qmail 67795 invoked by uid 60001); 1 Nov 2010 16:08:18 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1288627698; bh=t/UsHSD6TgOglOMtQSx9n4VaiuDTAMwEFe/LZ0M/OkY=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=rwPWuWTNaNn45kUvB/uHUDyV5z0Yz7fNRifd4UnWOeBceSTD6W620di++zbdQTYt2/PRIz35NNxMXbO789zHDiMl76ZVUbWQq4zHXLYOmSLcMTmKOhszuSeDEXI/0DFMWotWfIvLGPjwd2UvOoG1UNRVku2oJvHDHB9dVZFfcm0= DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=fAhLW2dWBRuNHfM8zRiSWi49DxCz9de9Fakl2ex8opgwvuGwZwXgNlBPcJ/4cGu/CJ6ZNHPTRUdA0B/0IbuutIsWEmHqqWb81euToemju+bEs5ncgr+18HO/hk7LxYDe1vV2VEIfmbW4hBkyc+Zq6eBvpVhmW3kJ+mmevZJOlO8=; Message-ID: <151658.65855.qm@web120002.mail.ne1.yahoo.com> X-YMail-OSG: dkRMJq8VM1mRod0qM0LfmD9L1dZAHPstaaig97htbEdR3Q7 pPYTciNxOjfVkEGGivMZ.w4c6z7.E997wSWxfjtBugnvqC95oKaWSQeVv4nH yOh.0yjUajQ_PijX.RcZBSOY.T4IMmrWotD7WvTGupkutE8r3aHwAPi6hrAV R2WhoaMnP02pATpVZijIV3xwYT80TmhqvYMC0 Received: from [68.36.212.155] by web120002.mail.ne1.yahoo.com via HTTP; Mon, 01 Nov 2010 09:08:18 PDT X-Mailer: YahooMailRC/504.5 YahooMailWebService/0.8.107.284920 Date: Mon, 1 Nov 2010 09:08:18 -0700 (PDT) From: A J Subject: Pagesize for large-objects (ONLY) database. To: pgsql-admin@postgresql.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-1408854558-1288627698=:65855" X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.971 tagged_above=-10 required=5 tests=BAYES_40=-0.001, HTML_FONT_FACE_BAD=0.981, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_NONE=-0.0001, T_RP_MATCHES_RCVD=-0.01 X-Spam-Level: X-Archive-Number: 201011/4 X-Sequence-Number: 35970 --0-1408854558-1288627698=:65855 Content-Type: text/plain; charset=us-ascii Hello, I am planning to have a postgres database for large object storage. I believe that with version 9.0, the 'world readable' issue with pg_largeobject is resolved and we can now control at object level- which users can read/write what objects. Streaming performance is an issue and hence will need to increase the PAGE SIZE (BLCKSZ ) to ensure access is sequential. Does anyone have recommendation on a BLCKSZ I can start with. There will be lot of reads,updates,inserts,deletes of the objects. Also any ideas on how big of an issue fragmentation can be and how can it be contained ? About 70% of the objects will be less than 1MB, about 25% upto 10MB and rest 5% or so may be larger than 10MB. Metadata management is also important so just a distributed filesystem (rather than database) is not viable. Any other suggestions are welcome. Thank you! - AJ --0-1408854558-1288627698=:65855 Content-Type: text/html; charset=us-ascii
Hello,
I am planning to have a postgres database for large object storage. I believe that with version 9.0, the 'world readable' issue with pg_largeobject is resolved and we can now control at object level- which users can read/write what objects.

Streaming performance is an issue and hence will need to increase the PAGE SIZE (BLCKSZ ) to ensure access is sequential. Does anyone have recommendation on a BLCKSZ I can start with.
There will be lot of reads,updates,inserts,deletes of the objects. Also any ideas on how big of an issue fragmentation can be and how can it be contained ?

About 70% of the objects will be less than 1MB, about 25% upto 10MB and rest 5% or so may be larger than 10MB.

Metadata management is also important so just a distributed filesystem (rather than database) is not viable. Any other suggestions are welcome.

Thank you! - AJ

--0-1408854558-1288627698=:65855--