Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f3O9p5f08435 for linux-xfs-outgoing; Tue, 24 Apr 2001 02:51:05 -0700 Received: from smtp.euronet.nl (america.euronet.nl [194.134.0.151]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f3O9p4M08431 for ; Tue, 24 Apr 2001 02:51:04 -0700 Received: from lowlands.euro.net (lowlands.euronet.nl [194.134.32.225]) by smtp.euronet.nl (Postfix) with ESMTP id 80AC420805 for ; Tue, 24 Apr 2001 11:51:02 +0200 (MET DST) X-NCC-RegID: nl.euronet Message-Id: <4.3.2.7.2.20010424113432.00bd2380@pop.euronet.nl> X-Sender: sengaia@pop.euronet.nl X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Tue, 24 Apr 2001 11:52:38 +0200 To: linux-xfs@oss.sgi.com From: Arjen Wolfs Subject: performance and hanging issues Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-linux-xfs@oss.sgi.com Precedence: bulk Hi, Since I only just subscribed here this might have been discussed before, but browsing through the archives I couldn't find anything about it so here goes: I am running 2.4.2-XFS (checked out of cvs) on a SMP box with 1G of memory, and have am having some serious issues with it. When creating a large number of files (a few thousand) on the XFS parition, the process will run along nicely for a while and then slow to a complete crawl; the process goes into an uninterruptable sleep (D), and the load goes up to ~1,5 with no process using any cpu. It might be stuck in this state for a minute or two, and then continue as normal for a while before the processes repeats itself. After about 3 days of uptime the box came to a complete halt, forcing a hard reset. It had the following messaged logged (though not at the time it froze): Apr 24 05:54:31 kernel: __alloc_pages: 0-order allocation failed. Apr 24 05:54:31 kernel: __alloc_pages: 0-order allocation failed. Apr 24 10:58:30 kernel: __alloc_pages: 0-order allocation failed. Apr 24 10:58:30 last message repeated 11 times The box froze about 40 minutes after that last log entry above. I know this all isn't very scientific and I can obviously attempt to provide more information. Any help would be extremely very greatly appreciated, /Arjen Wolfs