From jlan@sgi.com Fri Oct 1 17:42:18 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 01 Oct 2004 17:42:23 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i920gIMP013392 for ; Fri, 1 Oct 2004 17:42:18 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i921ql43025623 for ; Fri, 1 Oct 2004 18:52:47 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i920fxKh48622676; Fri, 1 Oct 2004 17:42:04 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i920cc88014749; Fri, 1 Oct 2004 17:38:48 -0700 Message-ID: <415DF88E.4020002@sgi.com> Date: Fri, 01 Oct 2004 17:38:38 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Paul Jackson CC: linux-kernel@vger.kernel.org, lse-tech@lists.sourceforge.net, csa@oss.sgi.com, akpm@osdl.org, guillaume.thouvenin@bull.net, tim@physik3.uni-rostock.de, corliss@digitalmages.com Subject: Re: [Lse-tech] Re: [PATCH 2.6.9-rc2 2/2] enhanced MM accounting data collection References: <4158956F.3030706@engr.sgi.com> <41589927.5080803@engr.sgi.com> <20040928023350.611c84d8.pj@sgi.com> In-Reply-To: <20040928023350.611c84d8.pj@sgi.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 40 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Paul Jackson wrote: > nits: > > 1) I'm not sure the "no-op if CONFIG_CSA not set" comments > are worthwhile - it does not seem to be a common practice > to mark macros that collapse under certain CONFIG's with > such comments, and some code, such as in fork.c, would > become quite a bit less readable if such comments were > widely used. Yeah, that makes sense. Will be fixed in next posting. > > 2) Three of the added csa_update_integrals() lines have > leading spaces, instead of a tab char, such as in: > > =================================================================== > --- linux.orig/fs/exec.c 2004-09-27 11:57:40.201435722 -0700 > +++ linux/fs/exec.c 2004-09-27 14:05:41.266160725 -0700 > @@ -1163,6 +1164,9 @@ > > /* execve success */ > security_bprm_free(&bprm); > + /* no-op if CONFIG_CSA not set */ > + csa_update_integrals(); <========= > + update_mem_hiwater(); <========= > return retval; > } Caused by 'cut-n-paste'. Will be fixed. > > 3) Is it always the case that csa_update_integrals() and > update_mem_hiwater() are used together? If so, perhaps > they could be collapsed into one? Even the current->mm > test inside them could be made one test, perhaps? As Robin pointed out, there are a couple of instances that are not the case. Actually there are three. Thanks for your feedback, Paul! - jay From fant@pobox.com Thu Oct 14 08:42:18 2004 Received: with ECARTIS (v1.0.0; list csa); Thu, 14 Oct 2004 08:42:23 -0700 (PDT) Received: from fuse1.fusemail.net (smtp.fusemail.net [69.31.1.141]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9EFgInM003236 for ; Thu, 14 Oct 2004 08:42:18 -0700 Received: from fusemail.com by fuse1.fusemail.net with asmtp (FuseMail extSMTP) id 1CI7jq-000446-I6 for csa@oss.sgi.com; Thu, 14 Oct 2004 10:41:58 -0500 Date: Thu, 14 Oct 2004 11:41:54 -0400 From: Andrew Fant Reply-To: Andrew Fant To: csa@oss.sgi.com Subject: nice missing from csa structures? Message-ID: <48530000.1097768514@flux.usg.tufts.edu> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 41 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: fant@pobox.com Precedence: bulk X-list: csa I'm attempting to build csa into a 2.4.26 kernel with the patches found on oss.sgi.com and have run into a strange problem. When make bzImage attempts to compile kernel/csa.c I get errors about nice not being defined in the csa structures. In particular, the lines: csa->ac_nice = p->nice; and eoj.ac_nice = current->nice; cause the compilation to fail. If I comment out those lines everything compiles and runs fine. Has anyone else had this issue, and has anyone found a workaround? I can't switch to a 2.6 kernel series, so that isn't an issue. On an unrelated note, can BSD process accounting and CSA coexist, or should I turn one off if the other is in use? Thanks, Andy From holt@lnx-holt.americas.sgi.com Thu Oct 14 08:48:25 2004 Received: with ECARTIS (v1.0.0; list csa); Thu, 14 Oct 2004 08:48:30 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9EFmPh1003425 for ; Thu, 14 Oct 2004 08:48:25 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [192.48.203.135]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9EH0jLi028122 for ; Thu, 14 Oct 2004 10:00:45 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9EFmAOV49325560; Thu, 14 Oct 2004 10:48:10 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9EFmAtC13035260; Thu, 14 Oct 2004 10:48:10 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9EFm9ej021976; Thu, 14 Oct 2004 10:48:09 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9EFm9Af021975; Thu, 14 Oct 2004 10:48:09 -0500 Date: Thu, 14 Oct 2004 10:48:09 -0500 From: Robin Holt To: Andrew Fant Cc: csa@oss.sgi.com Subject: Re: nice missing from csa structures? Message-ID: <20041014154809.GA21697@lnx-holt.americas.sgi.com> References: <48530000.1097768514@flux.usg.tufts.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <48530000.1097768514@flux.usg.tufts.edu> User-Agent: Mutt/1.4.1i X-archive-position: 42 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Thu, Oct 14, 2004 at 11:41:54AM -0400, Andrew Fant wrote: > I'm attempting to build csa into a 2.4.26 kernel with the patches found on > oss.sgi.com and have run into a strange problem. When make bzImage > attempts to compile kernel/csa.c I get errors about nice not being defined > in the csa structures. In particular, the lines: > > csa->ac_nice = p->nice; > > and > > eoj.ac_nice = current->nice; The nice field went away in one of the later 2.4 kernels. It is recorded in the eoj accounting record. If you are not concerned with filtering based upon process niceness, then you should be able to get by with eoj.ac_nice = 0; This will ensure you are not recording invalid data. Alternatively, you could look at the 2.6 series of patches and backport how it was done there. I believe you need to change it to something with task_nice(current), but it has been a long time since I looked at that code. Good Luck, Robin From j.logsdon@quantex-research.com Fri Oct 15 05:52:28 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 05:52:33 -0700 (PDT) Received: from heisenberg.zen.co.uk (heisenberg.zen.co.uk [212.23.3.141]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FCqRFS010113 for ; Fri, 15 Oct 2004 05:52:28 -0700 Received: from [217.155.43.225] (helo=quantex-research.com) by heisenberg.zen.co.uk with esmtp (Exim 4.30) id 1CIRZ7-0005xy-61 for csa@oss.sgi.com; Fri, 15 Oct 2004 12:52:13 +0000 Received: from localhost (j.logsdon@localhost) by quantex-research.com (8.8.7/8.8.7) with ESMTP id NAA18782 for ; Fri, 15 Oct 2004 13:52:13 +0100 Date: Fri, 15 Oct 2004 13:52:13 +0100 (GMT) From: John Logsdon X-Sender: j.logsdon@mercury.quantex To: csa@oss.sgi.com Subject: Availability Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Originating-Heisenberg-IP: [217.155.43.225] X-archive-position: 43 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: j.logsdon@quantex-research.com Precedence: bulk X-list: csa Hi all CSA looks a very interesting improvement on basic BSD accounting and many thanks to SGI for supporting it. Before I start trying it, I am working on 2.4.27. Is there a 2.4.27 patch upcoming - or can the 2.4.26 patch be used with minimal or even no hunks? The samre remarks for PAGG - which I gather you need for CSA anyway or are all the patches required in CSA? And which PAGG patch anyway? Best wishes John John Logsdon "Try to make things as simple Quantex Research Ltd, Manchester UK as possible but not simpler" j.logsdon@quantex-research.com a.einstein@relativity.org +44(0)161 445 4951/G:+44(0)7717758675 www.quantex-research.com From holt@lnx-holt.americas.sgi.com Fri Oct 15 06:14:57 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 06:15:03 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FDEu3k021094 for ; Fri, 15 Oct 2004 06:14:57 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FEROw3001456 for ; Fri, 15 Oct 2004 07:27:24 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FDCHJV005257; Fri, 15 Oct 2004 08:12:17 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FDCGtC12520963; Fri, 15 Oct 2004 08:12:16 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9FDCGiP000676; Fri, 15 Oct 2004 08:12:16 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9FDCGXQ000675; Fri, 15 Oct 2004 08:12:16 -0500 Date: Fri, 15 Oct 2004 08:12:16 -0500 From: Robin Holt To: John Logsdon Cc: csa@oss.sgi.com Subject: Re: Availability Message-ID: <20041015131216.GA633@lnx-holt.americas.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.1i X-archive-position: 44 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Fri, Oct 15, 2004 at 01:52:13PM +0100, John Logsdon wrote: > Hi all > > CSA looks a very interesting improvement on basic BSD accounting and many > thanks to SGI for supporting it. What are you trying to accomplish with your accounting. One of the key differences between csa and BSD accounting is job based accounting. Is that what you are trying to accomplish? > > Before I start trying it, I am working on 2.4.27. Is there a 2.4.27 patch > upcoming - or can the 2.4.26 patch be used with minimal or even no hunks? Unfortunately, I can't help you here. I haven't used csa for some time. In the distant past, you needed to apply pagg, job, and csa patches. I assume that seperation still applies. > > The samre remarks for PAGG - which I gather you need for CSA anyway or are > all the patches required in CSA? And which PAGG patch anyway? If you get it working, could you reply to the list with what you needed to do? If you get stuck, I can probably lend you a hand over the weekend. Good Luck, Robin Holt From erikj@subway.americas.sgi.com Fri Oct 15 06:30:34 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 06:30:40 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FDUYF2021399 for ; Fri, 15 Oct 2004 06:30:34 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FEh2D0006716 for ; Fri, 15 Oct 2004 07:43:02 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FDU9JV006357; Fri, 15 Oct 2004 08:30:10 -0500 (CDT) Received: from subway.americas.sgi.com (subway.americas.sgi.com [128.162.236.152]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FDU9tC13015449; Fri, 15 Oct 2004 08:30:09 -0500 (CDT) Received: from subway.americas.sgi.com (localhost [127.0.0.1]) by subway.americas.sgi.com (SGI-8.12.5/8.12.5/erikj-IRIX6519-news) with ESMTP id i9FDU9cc1205692; Fri, 15 Oct 2004 08:30:09 -0500 (CDT) Received: from localhost (erikj@localhost) by subway.americas.sgi.com (SGI-8.12.5/8.12.5/Submit) with ESMTP id i9FDU8Nn1205857; Fri, 15 Oct 2004 08:30:08 -0500 (CDT) Date: Fri, 15 Oct 2004 08:30:08 -0500 From: Erik Jacobson To: Robin Holt cc: John Logsdon , csa@oss.sgi.com Subject: Re: Availability In-Reply-To: <20041015131216.GA633@lnx-holt.americas.sgi.com> Message-ID: References: <20041015131216.GA633@lnx-holt.americas.sgi.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 45 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: erikj@subway.americas.sgi.com Precedence: bulk X-list: csa > > The samre remarks for PAGG - which I gather you need for CSA anyway or are > > all the patches required in CSA? And which PAGG patch anyway? > If you get it working, could you reply to the list with what you needed to > do? If you get stuck, I can probably lend you a hand over the weekend. I'm going to let Jay answer this one as he may know the best pairing. The PAGG patches for the most recent 2.6 kernels are a lot different than the 2.4 patches. If you were using fairly current 2.6 kernels, we'd just point you at the most recent PAGG and JOB patches. If it becomes necessary, we can look in to bringing the PAGG patch for 2.4 up to speed with what we're doing for 2.6. -- Erik Jacobson - Linux System Software - Silicon Graphics - Eagan, Minnesota From j.logsdon@quantex-research.com Fri Oct 15 06:54:31 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 06:54:39 -0700 (PDT) Received: from pythagoras.zen.co.uk (pythagoras.zen.co.uk [212.23.3.140]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FDsRCv027116 for ; Fri, 15 Oct 2004 06:54:28 -0700 Received: from [217.155.43.225] (helo=quantex-research.com) by pythagoras.zen.co.uk with esmtp (Exim 4.30) id 1CISX6-0004AP-Kg; Fri, 15 Oct 2004 13:54:12 +0000 Received: from localhost (j.logsdon@localhost) by quantex-research.com (8.8.7/8.8.7) with ESMTP id OAA01975; Fri, 15 Oct 2004 14:54:14 +0100 Date: Fri, 15 Oct 2004 14:54:13 +0100 (GMT) From: John Logsdon X-Sender: j.logsdon@mercury.quantex To: Erik Jacobson cc: Robin Holt , csa@oss.sgi.com Subject: Re: Availability In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Originating-Pythagoras-IP: [217.155.43.225] X-archive-position: 46 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: j.logsdon@quantex-research.com Precedence: bulk X-list: csa Erik, Robin, all Thanks for the immediate responses. I was just composing a reply to Robin's so I may as well put it all together. It is the job and user based accounting that is very interesting in CSA but other aspects of the implementation are more complete than the BSD version. I will see what happens... At the moment I have grsecurity-based accounting which is process based. This has a lot of advantages for looking at individual programs and reports elapsed and cpu times by UID, GID, EUID, EGID and parent process. I may still retain that - it is hooked in a completely different place to BSD so I suspect also to CSA so it shouldn't interfere. The current grsec kernel is 2.4.27 - there are one or two non-grsec security issues that fixes over 2.4.26. Other than the grsec parts, the kernel is vanilla and I compile it without modules. I don't want to use the 2.6 kernel as yet - while it has advantages for example in the scheduler (I am using a Xeon-based box) it is a little too early to use on a production system. Maybe you bleeding-edge guys will think differently but I have to be conservative in this application. So the real issue for me is 2.4.x and the .28 kernel is still in pre-release phase. Surely in time the 2.6 kernel will be fine. I have a little time to decide on this - process accounting is not required tomorrow but probably before 2.6 takes over completely from 2.4. Since CSA requires PAGG, in due course it would be appropriate to combine the patches but I guess you can always append the patch files and run it in one go. What's the JOB patch? I don't see it on the SGI project list. (Actually your ftp server seems to hang episodically, particularly on IE). Best wishes John John Logsdon "Try to make things as simple Quantex Research Ltd, Manchester UK as possible but not simpler" j.logsdon@quantex-research.com a.einstein@relativity.org +44(0)161 445 4951/G:+44(0)7717758675 www.quantex-research.com On Fri, 15 Oct 2004, Erik Jacobson wrote: > > > The samre remarks for PAGG - which I gather you need for CSA anyway or are > > > all the patches required in CSA? And which PAGG patch anyway? > > If you get it working, could you reply to the list with what you needed to > > do? If you get stuck, I can probably lend you a hand over the weekend. > > I'm going to let Jay answer this one as he may know the best pairing. The > PAGG patches for the most recent 2.6 kernels are a lot different than the > 2.4 patches. > > If you were using fairly current 2.6 kernels, we'd just point you at the > most recent PAGG and JOB patches. > > If it becomes necessary, we can look in to bringing the PAGG patch for 2.4 > up to speed with what we're doing for 2.6. > > -- > Erik Jacobson - Linux System Software - Silicon Graphics - Eagan, Minnesota > From erikj@subway.americas.sgi.com Fri Oct 15 07:05:58 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 07:06:03 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FE5vpm003202 for ; Fri, 15 Oct 2004 07:05:58 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FFIQ2F017008 for ; Fri, 15 Oct 2004 08:18:26 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FE3IJV008407; Fri, 15 Oct 2004 09:03:18 -0500 (CDT) Received: from subway.americas.sgi.com (subway.americas.sgi.com [128.162.236.152]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FE3ItC13098436; Fri, 15 Oct 2004 09:03:18 -0500 (CDT) Received: from subway.americas.sgi.com (localhost [127.0.0.1]) by subway.americas.sgi.com (SGI-8.12.5/8.12.5/erikj-IRIX6519-news) with ESMTP id i9FE3Hcc1205969; Fri, 15 Oct 2004 09:03:17 -0500 (CDT) Received: from localhost (erikj@localhost) by subway.americas.sgi.com (SGI-8.12.5/8.12.5/Submit) with ESMTP id i9FE3HFr1204047; Fri, 15 Oct 2004 09:03:17 -0500 (CDT) Date: Fri, 15 Oct 2004 09:03:17 -0500 From: Erik Jacobson To: John Logsdon cc: Robin Holt , csa@oss.sgi.com Subject: Re: Availability In-Reply-To: Message-ID: References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 47 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: erikj@subway.americas.sgi.com Precedence: bulk X-list: csa I read your not. I'm still going to see what Jay says about this. If he thinks we need to bring the PAGG and JOB patches up to date for 2.4.x, I'd be happy to work on that. > Since CSA requires PAGG, in due course it would be appropriate to combine > the patches but I guess you can always append the patch files and run it > in one go. What's the JOB patch? I don't see it on the SGI project list. > (Actually your ftp server seems to hang episodically, particularly on IE). I don't know anything about the web site hanging - sorry. But here is the project page for PAGG and Job: http://oss.sgi.com/projects/pagg/ PAGG is sort of a generic patch for making kernel modules that need to group processes together. It provides callbacks that can notify your kernel module when a process forks, execs, exits, etc. It is not specific to CSA - we have things at SGI that make use of PAGG but not CSA and Job. Job provides inescapable job containers. It is one of the users of PAGG. CSA uses both PAGG and Job. Job on it's own could be useful, for example, for people writing batch schedulers, where you want to track a job container that may contain multiple processes and manage it (throttle it, etc). Since a given job may need to start other processes to do it's work, it is convenient to be able to collect it together inside a container even if you're not interested in running CSA. I guess I'm trying to say that Job and especially PAGG stand on their own for the most part... I hope that helps. Erik From andrew.fant@tufts.edu Fri Oct 15 07:52:59 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 07:53:04 -0700 (PDT) Received: from andesite.usg.tufts.edu (andesite.usg.tufts.edu [130.64.1.202]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FEqwAv004287 for ; Fri, 15 Oct 2004 07:52:59 -0700 Received: from flux.usg.tufts.edu ([130.64.100.43]) by andesite.usg.tufts.edu with esmtp (Exim 4.20) id 1CITRk-0003tp-k2 for csa@oss.sgi.com; Fri, 15 Oct 2004 10:52:44 -0400 Date: Fri, 15 Oct 2004 10:52:44 -0400 From: Andrew Fant To: csa@oss.sgi.com Subject: Two Simple Questions Message-ID: <31770000.1097851964@flux.usg.tufts.edu> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 48 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: andrew.fant@tufts.edu Precedence: bulk X-list: csa Wow, activity on the mailing list just as I get CSA installed on my testbed. My timing gets better and better. I have two questions for anyone who might have answers 1) Has anyone gotten CSA to work with LSF under Linux? 2) When I use ja, all my reports terminate with a segfault. For example: afant@chicken04% ja -s Job CSA Accounting - Summary Report ==================================== Job Accounting File Name : /var/tmp/.jaccta8c04406000063b3 Operating System : Linux chicken04 2.4.26-gentoo-r9 #4 SMP Thu Oct 14 13:35:37 EDT 2004 i686 User Name (ID) : afant (1018) Group Name (ID) : tccs (499) Project Name (ID) : ? (0) Job ID : 0xa8c04406000063b3 Report Starts : 10/15/04 10:48:16 Report Ends : 10/15/04 10:50:44 Elapsed Time : 148 Seconds User CPU Time : 20.0000 Seconds System CPU Time : 7.3600 Seconds Block I/O Wait Time : 0.0000 Seconds Raw I/O Wait Time : 0.0000 Seconds CPU Time Core Memory Integral : 17592190.1377 Mbyte-seconds CPU Time Virtual Memory Integral : 52776541.9053 Mbyte-seconds Maximum Core Memory Used : 99.9805 Mbytes Maximum Virtual Memory Used : 159.7500 Mbytes Characters Read : 286.7011 Mbytes Characters Written : 289.2352 Mbytes Blocks Read : 0 Blocks Written : 0 Logical I/O Read Requests : 6668 Logical I/O Write Requests : 5173 Number of Commands : 11 System Billing Units : 0.0000 Segmentation fault Has anyone else ever seen this behavior? Thanks, Andy From holt@lnx-holt.americas.sgi.com Fri Oct 15 07:55:15 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 07:55:20 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FEtF8H004353 for ; Fri, 15 Oct 2004 07:55:15 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FG7hXM031182 for ; Fri, 15 Oct 2004 09:07:43 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FEsoJV011489; Fri, 15 Oct 2004 09:54:50 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FEsmtC13075409; Fri, 15 Oct 2004 09:54:48 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9FEsmio001474; Fri, 15 Oct 2004 09:54:48 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9FEsm2L001473; Fri, 15 Oct 2004 09:54:48 -0500 Date: Fri, 15 Oct 2004 09:54:48 -0500 From: Robin Holt To: John Logsdon Cc: Erik Jacobson , Robin Holt , csa@oss.sgi.com Subject: Re: Availability Message-ID: <20041015145448.GB873@lnx-holt.americas.sgi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.1i X-archive-position: 49 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Fri, Oct 15, 2004 at 02:54:13PM +0100, John Logsdon wrote: > Erik, Robin, all > > Thanks for the immediate responses. I was just composing a reply to > Robin's so I may as well put it all together. > > It is the job and user based accounting that is very interesting in CSA > but other aspects of the implementation are more complete than the BSD > version. I will see what happens... csa will record accounting stats on process exits, but only when the jobid is non-zero. You could hack that up quickly. So far, there have been a few requests to do that, but it never gets a lot of traction as it defeats the purpose of job based accounting. Usually, this type of accounting is used at large shops that want to attribute resource usage back to a customer and do not really care about system accounting. Thanks, Robin > > At the moment I have grsecurity-based accounting which is process based. > This has a lot of advantages for looking at individual programs and > reports elapsed and cpu times by UID, GID, EUID, EGID and parent process. > I may still retain that - it is hooked in a completely different place to > BSD so I suspect also to CSA so it shouldn't interfere. The current grsec > kernel is 2.4.27 - there are one or two non-grsec security issues that > fixes over 2.4.26. Other than the grsec parts, the kernel is vanilla and > I compile it without modules. > > I don't want to use the 2.6 kernel as yet - while it has advantages for > example in the scheduler (I am using a Xeon-based box) it is a little too > early to use on a production system. Maybe you bleeding-edge guys will > think differently but I have to be conservative in this application. So > the real issue for me is 2.4.x and the .28 kernel is still in pre-release > phase. Surely in time the 2.6 kernel will be fine. I have a little time > to decide on this - process accounting is not required tomorrow but > probably before 2.6 takes over completely from 2.4. > > Since CSA requires PAGG, in due course it would be appropriate to combine > the patches but I guess you can always append the patch files and run it > in one go. What's the JOB patch? I don't see it on the SGI project list. > (Actually your ftp server seems to hang episodically, particularly on IE). > > Best wishes > > John > > John Logsdon "Try to make things as simple > Quantex Research Ltd, Manchester UK as possible but not simpler" > j.logsdon@quantex-research.com a.einstein@relativity.org > +44(0)161 445 4951/G:+44(0)7717758675 www.quantex-research.com > > > On Fri, 15 Oct 2004, Erik Jacobson wrote: > > > > > The samre remarks for PAGG - which I gather you need for CSA anyway or are > > > > all the patches required in CSA? And which PAGG patch anyway? > > > If you get it working, could you reply to the list with what you needed to > > > do? If you get stuck, I can probably lend you a hand over the weekend. > > > > I'm going to let Jay answer this one as he may know the best pairing. The > > PAGG patches for the most recent 2.6 kernels are a lot different than the > > 2.4 patches. > > > > If you were using fairly current 2.6 kernels, we'd just point you at the > > most recent PAGG and JOB patches. > > > > If it becomes necessary, we can look in to bringing the PAGG patch for 2.4 > > up to speed with what we're doing for 2.6. > > > > -- > > Erik Jacobson - Linux System Software - Silicon Graphics - Eagan, Minnesota > > > > From holt@lnx-holt.americas.sgi.com Fri Oct 15 08:23:32 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 08:23:36 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FFNVeR008325 for ; Fri, 15 Oct 2004 08:23:32 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FGa0ad007566 for ; Fri, 15 Oct 2004 09:36:00 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FFKoJV013180; Fri, 15 Oct 2004 10:20:50 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FFKntC12971076; Fri, 15 Oct 2004 10:20:50 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9FFKnDn001700; Fri, 15 Oct 2004 10:20:49 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9FFKnLC001699; Fri, 15 Oct 2004 10:20:49 -0500 Date: Fri, 15 Oct 2004 10:20:49 -0500 From: Robin Holt To: Andrew Fant Cc: csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <20041015152049.GC873@lnx-holt.americas.sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <31770000.1097851964@flux.usg.tufts.edu> User-Agent: Mutt/1.4.1i X-archive-position: 50 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Fri, Oct 15, 2004 at 10:52:44AM -0400, Andrew Fant wrote: > Wow, activity on the mailing list just as I get CSA installed on my > testbed. My timing gets better and better. I have two questions for > anyone who might have answers > > 1) Has anyone gotten CSA to work with LSF under Linux? We had it working with lsf and PBS Pro under the 2.4 kernel. I am not sure if it is working there now. > > 2) When I use ja, all my reports terminate with a segfault. For example: Do you have core dumps turned on? If so, you might want to gdb the core dump and find out why or at least which function you are in. System Billing Units is the last thing that should be output. Good Luck, Robin From jlan@sgi.com Fri Oct 15 10:01:26 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 10:01:31 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FH1QSL018604 for ; Fri, 15 Oct 2004 10:01:26 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FIDt9l002995 for ; Fri, 15 Oct 2004 11:13:55 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i9FH06Kh52124719; Fri, 15 Oct 2004 10:00:11 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i9FH140V002772; Fri, 15 Oct 2004 10:01:25 -0700 Message-ID: <41700250.6060800@sgi.com> Date: Fri, 15 Oct 2004 10:01:04 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Andrew Fant CC: csa@oss.sgi.com Subject: Re: nice missing from csa structures? References: <48530000.1097768514@flux.usg.tufts.edu> In-Reply-To: <48530000.1097768514@flux.usg.tufts.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 51 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Sorry for being slow to respond. My email filter did not set up to allow in emails from oss.sgi.com :( Problem is corrected. Robin answered your question on nice feature. If your kernel/sched.c contains task_nice routine, you may fix the problem as below: csa->ac_nice = task_nice(p); and eoj.ac_nice = task_nice(current); As to BSD accounting, CSA can coexist with it. You do not need to turn CONFIG_BSD_PROCESS_ACCT off. Thanks! - jay Andrew Fant wrote: > I'm attempting to build csa into a 2.4.26 kernel with the patches found > on oss.sgi.com and have run into a strange problem. When make bzImage > attempts to compile kernel/csa.c I get errors about nice not being > defined in the csa structures. In particular, the lines: > > csa->ac_nice = p->nice; > > and > > eoj.ac_nice = current->nice; > > cause the compilation to fail. If I comment out those lines everything > compiles and runs fine. Has anyone else had this issue, and has anyone > found a workaround? I can't switch to a 2.6 kernel series, so that > isn't an issue. > > On an unrelated note, can BSD process accounting and CSA coexist, or > should I turn one off if the other is in use? > > Thanks, > Andy > From jlan@sgi.com Fri Oct 15 10:29:19 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 10:29:24 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FHTJWC020972 for ; Fri, 15 Oct 2004 10:29:19 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FIfm9x010885 for ; Fri, 15 Oct 2004 11:41:48 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i9FHSvKh49835733; Fri, 15 Oct 2004 10:29:02 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i9FHTj0V002788; Fri, 15 Oct 2004 10:30:05 -0700 Message-ID: <41700909.1020308@sgi.com> Date: Fri, 15 Oct 2004 10:29:45 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Erik Jacobson CC: John Logsdon , Robin Holt , csa@oss.sgi.com Subject: Re: Availability References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 52 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Hi, Our current published 2.4 patchset for pagg/job/csa is at 2.4.26. Please let us know if these patches causes conflict and are not obvious to fix, we will release the patchset for 2.4.27. The csa 2.1.x rpm works with 2.4.2.[25,26] kernel patchset as well as with 2.6 kernel patches. You can get job patch from the same place as the pagg patch. Although pagg and job can stand alone as Erik stated, CSA needs to have pagg and job kernel patches and the job rpm to be fully functional. A word on 2.6 CSA kernel. I broke the single 2.4 CSA kernel patches into several pieces. We tried to integrate the accounting data collection in the area of I/O and MM into general accounting so that people using BSD, ELSA or CSA can all benefit. Keep in mind that the 2.4 CSA patch is still in the one big piece form. Thanks! - jay Erik Jacobson wrote: > I read your not. I'm still going to see what Jay says about this. If > he thinks we need to bring the PAGG and JOB patches up to date for 2.4.x, > I'd be happy to work on that. > > >>Since CSA requires PAGG, in due course it would be appropriate to combine >>the patches but I guess you can always append the patch files and run it >>in one go. What's the JOB patch? I don't see it on the SGI project list. >>(Actually your ftp server seems to hang episodically, particularly on IE). > > > I don't know anything about the web site hanging - sorry. > > But here is the project page for PAGG and Job: > http://oss.sgi.com/projects/pagg/ > > PAGG is sort of a generic patch for making kernel modules that need to > group processes together. It provides callbacks that can notify your > kernel module when a process forks, execs, exits, etc. It is not > specific to CSA - we have things at SGI that make use of PAGG but not > CSA and Job. > > Job provides inescapable job containers. It is one of the users of PAGG. > CSA uses both PAGG and Job. > > Job on it's own could be useful, for example, for people writing batch > schedulers, where you want to track a job container that may contain multiple > processes and manage it (throttle it, etc). Since a given job may need to > start other processes to do it's work, it is convenient to be able to > collect it together inside a container even if you're not interested in > running CSA. > > I guess I'm trying to say that Job and especially PAGG stand on their own > for the most part... > > I hope that helps. > > Erik From jlan@sgi.com Fri Oct 15 10:37:47 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 10:37:51 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FHbk9o021113 for ; Fri, 15 Oct 2004 10:37:46 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FIoGj7013154 for ; Fri, 15 Oct 2004 11:50:16 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i9FHbRKh52004028; Fri, 15 Oct 2004 10:37:32 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i9FHcP0V006325; Fri, 15 Oct 2004 10:38:35 -0700 Message-ID: <41700B10.1070102@sgi.com> Date: Fri, 15 Oct 2004 10:38:24 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Robin Holt CC: Andrew Fant , csa@oss.sgi.com Subject: Re: Two Simple Questions References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> In-Reply-To: <20041015152049.GC873@lnx-holt.americas.sgi.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 53 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Robin Holt wrote: > On Fri, Oct 15, 2004 at 10:52:44AM -0400, Andrew Fant wrote: > >>Wow, activity on the mailing list just as I get CSA installed on my >>testbed. My timing gets better and better. I have two questions for >>anyone who might have answers >> >>1) Has anyone gotten CSA to work with LSF under Linux? > > > We had it working with lsf and PBS Pro under the 2.4 kernel. I > am not sure if it is working there now. Yes, still working! :) > > >>2) When I use ja, all my reports terminate with a segfault. For example: Did you install job rpm, chkconfig on job, and modify /etc/pam.d/ files as noted when you installed job rpm? I have not received reports on ja segfault before. Thanks, - jay > > > Do you have core dumps turned on? If so, you might want to gdb > the core dump and find out why or at least which function you > are in. > > System Billing Units is the last thing that should be output. > > Good Luck, > Robin From holt@lnx-holt.americas.sgi.com Fri Oct 15 10:41:25 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 10:41:29 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FHfOZi021199 for ; Fri, 15 Oct 2004 10:41:25 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FIrs7K014040 for ; Fri, 15 Oct 2004 11:53:54 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FHcdJV020946; Fri, 15 Oct 2004 12:38:39 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FHcbtC13049143; Fri, 15 Oct 2004 12:38:38 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9FHcbr6002762; Fri, 15 Oct 2004 12:38:37 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9FHcbGu002760; Fri, 15 Oct 2004 12:38:37 -0500 Date: Fri, 15 Oct 2004 12:38:37 -0500 From: Robin Holt To: Jay Lan Cc: Robin Holt , Andrew Fant , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <20041015173837.GG873@lnx-holt.americas.sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <41700B10.1070102@sgi.com> User-Agent: Mutt/1.4.1i X-archive-position: 54 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Fri, Oct 15, 2004 at 10:38:24AM -0700, Jay Lan wrote: > Robin Holt wrote: > >On Fri, Oct 15, 2004 at 10:52:44AM -0400, Andrew Fant wrote: > > > >>Wow, activity on the mailing list just as I get CSA installed on my > >>testbed. My timing gets better and better. I have two questions for > >>anyone who might have answers > >> > >>1) Has anyone gotten CSA to work with LSF under Linux? > > > > > >We had it working with lsf and PBS Pro under the 2.4 kernel. I > >am not sure if it is working there now. > > Yes, still working! :) > > > > > > >>2) When I use ja, all my reports terminate with a segfault. For example: > > Did you install job rpm, chkconfig on job, and modify /etc/pam.d/ files > as noted when you installed job rpm? > > I have not received reports on ja segfault before. I don't think you can get the job accounting started without the above. Let me try. Robin From holt@lnx-holt.americas.sgi.com Fri Oct 15 10:45:42 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 10:45:47 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FHjgv8021319 for ; Fri, 15 Oct 2004 10:45:42 -0700 Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [198.149.16.15]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FIwBuY015201 for ; Fri, 15 Oct 2004 11:58:11 -0700 Received: from thistle-e236.americas.sgi.com (thistle-e236.americas.sgi.com [128.162.236.204]) by flecktone.americas.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9FHguJV021221; Fri, 15 Oct 2004 12:42:56 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (lnx-holt.americas.sgi.com [128.162.233.109]) by thistle-e236.americas.sgi.com (8.12.9/SGI-server-1.8) with ESMTP id i9FHgstC13079960; Fri, 15 Oct 2004 12:42:55 -0500 (CDT) Received: from lnx-holt.americas.sgi.com (localhost.localdomain [127.0.0.1]) by lnx-holt.americas.sgi.com (8.12.11/8.12.11) with ESMTP id i9FHgsPH002800; Fri, 15 Oct 2004 12:42:54 -0500 Received: (from holt@localhost) by lnx-holt.americas.sgi.com (8.12.11/8.12.11/Submit) id i9FHgs72002799; Fri, 15 Oct 2004 12:42:54 -0500 Date: Fri, 15 Oct 2004 12:42:54 -0500 From: Robin Holt To: Robin Holt Cc: Jay Lan , Andrew Fant , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <20041015174254.GH873@lnx-holt.americas.sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20041015173837.GG873@lnx-holt.americas.sgi.com> User-Agent: Mutt/1.4.1i X-archive-position: 55 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: holt@sgi.com Precedence: bulk X-list: csa On Fri, Oct 15, 2004 at 12:38:37PM -0500, Robin Holt wrote: > On Fri, Oct 15, 2004 at 10:38:24AM -0700, Jay Lan wrote: > > Robin Holt wrote: > > >On Fri, Oct 15, 2004 at 10:52:44AM -0400, Andrew Fant wrote: > > > > > >>Wow, activity on the mailing list just as I get CSA installed on my > > >>testbed. My timing gets better and better. I have two questions for > > >>anyone who might have answers > > >> > > >>1) Has anyone gotten CSA to work with LSF under Linux? > > > > > > > > >We had it working with lsf and PBS Pro under the 2.4 kernel. I > > >am not sure if it is working there now. > > > > Yes, still working! :) > > > > > > > > > > >>2) When I use ja, all my reports terminate with a segfault. For example: > > > > Did you install job rpm, chkconfig on job, and modify /etc/pam.d/ files > > as noted when you installed job rpm? > > > > I have not received reports on ja segfault before. > > I don't think you can get the job accounting started without the above. > Let me try. Nope, there must be something else going on. We would need to know a lot more to help you. I guess we would need to start with kernel version and patches, job and csa userland versions, glibc version, and compiler version. It will probably be easier to get a core dump and just issue 'where'. That will probably be the best information. Thanks, Robin From andrew.fant@tufts.edu Fri Oct 15 12:01:08 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 12:01:14 -0700 (PDT) Received: from andesite.usg.tufts.edu (andesite.usg.tufts.edu [130.64.1.202]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FJ18Xq024434 for ; Fri, 15 Oct 2004 12:01:08 -0700 Received: from flux.usg.tufts.edu ([130.64.100.43]) by andesite.usg.tufts.edu with esmtp (Exim 4.20) id 1CIXJu-0007SQ-jR; Fri, 15 Oct 2004 15:00:54 -0400 Date: Fri, 15 Oct 2004 15:00:54 -0400 From: Andrew Fant To: Robin Holt cc: Jay Lan , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <90790000.1097866854@flux.usg.tufts.edu> In-Reply-To: <20041015174254.GH873@lnx-holt.americas.sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 56 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: andrew.fant@tufts.edu Precedence: bulk X-list: csa --On Friday, October 15, 2004 12:42:54 -0500 Robin Holt wrote: > On Fri, Oct 15, 2004 at 12:38:37PM -0500, Robin Holt wrote: >> On Fri, Oct 15, 2004 at 10:38:24AM -0700, Jay Lan wrote: >> > Robin Holt wrote: >> > > On Fri, Oct 15, 2004 at 10:52:44AM -0400, Andrew Fant wrote: >> > > >> > >> Wow, activity on the mailing list just as I get CSA installed on my >> > >> testbed. My timing gets better and better. I have two questions >> > >> for anyone who might have answers >> > >> >> > >> 1) Has anyone gotten CSA to work with LSF under Linux? >> > > >> > > >> > > We had it working with lsf and PBS Pro under the 2.4 kernel. I >> > > am not sure if it is working there now. >> > >> > Yes, still working! :) >> > >> > > >> > > >> > >> 2) When I use ja, all my reports terminate with a segfault. For >> > >> example: >> > >> > Did you install job rpm, chkconfig on job, and modify /etc/pam.d/ files >> > as noted when you installed job rpm? >> > >> > I have not received reports on ja segfault before. >> >> I don't think you can get the job accounting started without the above. >> Let me try. > > Nope, there must be something else going on. We would need to know > a lot more to help you. I guess we would need to start with kernel > version and patches, job and csa userland versions, glibc version, > and compiler version. > > It will probably be easier to get a core dump and just issue 'where'. > That will probably be the best information. > > Thanks, > Robin > > > Well, the core dump was less than helpful. The traceback I got was: (gdb) where #0 0x0e10567b in __register_atfork () from /lib/libc.so.6 #1 0x0e07199f in __cxa_finalize () from /lib/libc.so.6 #2 0x0e01f590 in ?? () from /lib/libm.so.6 #3 0x0e03d600 in ?? () from /lib/libm.so.6 #4 0x0e03d720 in ?? () from /lib/libm.so.6 #5 0xdfffeb18 in ?? () #6 0x0e0370d6 in ?? () from /lib/libm.so.6 #7 0x0e01c000 in ?? () #8 0x0e0139dc in ?? () from /lib/ld-linux.so.2 #9 0xdfffeba8 in ?? () #10 0x0e00afc6 in _dl_rtld_di_serinfo () from /lib/ld-linux.so.2 Previous frame inner to this frame (corrupt stack?) I'm going to try another kernel rebuild and see if that helps. Andy From andrew.fant@tufts.edu Fri Oct 15 13:21:21 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 13:21:25 -0700 (PDT) Received: from dacite.usg.tufts.edu (dacite.usg.tufts.edu [130.64.1.203]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FKLKVj000525 for ; Fri, 15 Oct 2004 13:21:20 -0700 Received: from flux.usg.tufts.edu ([130.64.100.43]) by dacite.usg.tufts.edu with esmtp (Exim 4.20) id 1CIYZW-0000ax-Zh; Fri, 15 Oct 2004 16:21:06 -0400 Date: Fri, 15 Oct 2004 16:21:06 -0400 From: Andrew Fant To: Robin Holt cc: Jay Lan , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <111360000.1097871666@flux.usg.tufts.edu> In-Reply-To: <20041015174254.GH873@lnx-holt.americas.sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 57 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: andrew.fant@tufts.edu Precedence: bulk X-list: csa --On Friday, October 15, 2004 12:42:54 -0500 Robin Holt wrote: > > Nope, there must be something else going on. We would need to know > a lot more to help you. I guess we would need to start with kernel > version and patches, job and csa userland versions, glibc version, > and compiler version. > > It will probably be easier to get a core dump and just issue 'where'. > That will probably be the best information. Robin: I do have job and csa set to run at boot-time, the pam.d files have been modified, jobtest runs successfully, and jstat does return a jid when I log in. I installed job and csa from sources, not from the rpm. Oddly, I have just discovered a second error mode: chicken04 root # /usr/sbin/csaswitch -c halt /proc/csa ioctl failure, command='csa_halt' System Error(14): Bad address. Unable to halt system accounting. System Error(14): Bad address. As for system specs, I have reproduced the error with both a vanilla un-patched 2.4.26 kernel and the patched gentoo-sources 2.4.26 kernel. Glibc is 2.3.3. gcc is version 3.3.3. I am using job 1.4 and csa userland 2.2.0. For pagg and job I am using the 2.4.26-4 patches. Thanks for your help so far. If there is anything else you want to know, please let me know. Andy From jlan@sgi.com Fri Oct 15 13:46:40 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 13:46:45 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FKkdo6000961 for ; Fri, 15 Oct 2004 13:46:40 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FLxApb003451 for ; Fri, 15 Oct 2004 14:59:10 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i9FKkIKh50226603; Fri, 15 Oct 2004 13:46:23 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i9FKlG0V013709; Fri, 15 Oct 2004 13:47:37 -0700 Message-ID: <41703754.2040409@sgi.com> Date: Fri, 15 Oct 2004 13:47:16 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Andrew Fant CC: Robin Holt , csa@oss.sgi.com Subject: Re: Two Simple Questions References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> <111360000.1097871666@flux.usg.tufts.edu> In-Reply-To: <111360000.1097871666@flux.usg.tufts.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 58 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Sorry about the csa_halt bug. It has been identified and fixed. We know that bug had nothing to do with the ja problem you mentioned. To get around the csa_halt problem, please just recycle the csa by doing "/etc/init.d/csa stop; /etc/init.d/csa start". Thanks, - jay Andrew Fant wrote: > > > --On Friday, October 15, 2004 12:42:54 -0500 Robin Holt > wrote: > >> >> Nope, there must be something else going on. We would need to know >> a lot more to help you. I guess we would need to start with kernel >> version and patches, job and csa userland versions, glibc version, >> and compiler version. >> >> It will probably be easier to get a core dump and just issue 'where'. >> That will probably be the best information. > > > Robin: > I do have job and csa set to run at boot-time, the pam.d files have > been modified, jobtest runs successfully, and jstat does return a jid > when I log in. I installed job and csa from sources, not from the rpm. > > Oddly, I have just discovered a second error mode: > > chicken04 root # /usr/sbin/csaswitch -c halt > /proc/csa ioctl failure, command='csa_halt' > System Error(14): Bad address. > Unable to halt system accounting. > System Error(14): Bad address. > > As for system specs, I have reproduced the error with both a vanilla > un-patched 2.4.26 kernel and the patched gentoo-sources 2.4.26 kernel. > Glibc is 2.3.3. gcc is version 3.3.3. I am using job 1.4 and csa > userland 2.2.0. For pagg and job I am using the 2.4.26-4 patches. > > Thanks for your help so far. If there is anything else you want to > know, please let me know. > > Andy From andrew.fant@tufts.edu Fri Oct 15 13:51:37 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 13:51:42 -0700 (PDT) Received: from basalt.usg.tufts.edu (basalt.usg.tufts.edu [130.64.1.201]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FKpbrY001062 for ; Fri, 15 Oct 2004 13:51:37 -0700 Received: from flux.usg.tufts.edu ([130.64.100.43]) by basalt.usg.tufts.edu with esmtp (Exim 4.20) id 1CIZ2p-0006o9-Q3; Fri, 15 Oct 2004 16:51:23 -0400 Date: Fri, 15 Oct 2004 16:51:23 -0400 From: Andrew Fant To: Jay Lan cc: Robin Holt , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <123050000.1097873483@flux.usg.tufts.edu> In-Reply-To: <41703754.2040409@sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> <111360000.1097871666@flux.usg.tufts.edu> <41703754.2040409@sgi.com> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 59 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: andrew.fant@tufts.edu Precedence: bulk X-list: csa --On Friday, October 15, 2004 13:47:16 -0700 Jay Lan wrote: > > Sorry about the csa_halt bug. It has been identified and fixed. > We know that bug had nothing to do with the ja problem you mentioned. > > To get around the csa_halt problem, please just recycle the csa > by doing "/etc/init.d/csa stop; /etc/init.d/csa start". > > Thanks, > - jay Jay, Thanks for the information, but /etc/init.d/csa stop calls /usr/sbin/csaswitch -c halt, so that won't actually stop it, as far as I can tell. Andy From jlan@sgi.com Fri Oct 15 14:10:16 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 14:10:21 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FLAGKu002048 for ; Fri, 15 Oct 2004 14:10:16 -0700 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9FMMlSq009323 for ; Fri, 15 Oct 2004 15:22:47 -0700 Received: from mtv-vpn-hw-jlan-2.corp.sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [134.15.18.195]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id i9FL7eKh50517747; Fri, 15 Oct 2004 14:07:45 -0700 (PDT) Received: from sgi.com (mtv-vpn-hw-jlan-2.corp.sgi.com [127.0.0.1]) by mtv-vpn-hw-jlan-2.corp.sgi.com (8.12.8/8.12.8) with ESMTP id i9FL8c0V013762; Fri, 15 Oct 2004 14:08:59 -0700 Message-ID: <41703C56.3040805@sgi.com> Date: Fri, 15 Oct 2004 14:08:38 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: zh-tw, en-us, en, zh-cn, zh-hk MIME-Version: 1.0 To: Andrew Fant CC: Robin Holt , csa@oss.sgi.com Subject: Re: Two Simple Questions References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> <111360000.1097871666@flux.usg.tufts.edu> <41703754.2040409@sgi.com> <123050000.1097873483@flux.usg.tufts.edu> In-Reply-To: <123050000.1097873483@flux.usg.tufts.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 60 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@sgi.com Precedence: bulk X-list: csa Yes, you were right. But then 'rmmod csa' cleans it up. Just a workaround until the real fix is released. I will try to release the fixes later next week. It needs fixes to both the rpm and kernel patch. - jay Andrew Fant wrote: > > > --On Friday, October 15, 2004 13:47:16 -0700 Jay Lan wrote: > >> >> Sorry about the csa_halt bug. It has been identified and fixed. >> We know that bug had nothing to do with the ja problem you mentioned. >> >> To get around the csa_halt problem, please just recycle the csa >> by doing "/etc/init.d/csa stop; /etc/init.d/csa start". >> >> Thanks, >> - jay > > > > Jay, > Thanks for the information, but /etc/init.d/csa stop calls > /usr/sbin/csaswitch -c halt, so that won't actually stop it, as far as I > can tell. > > Andy From andrew.fant@tufts.edu Fri Oct 15 14:20:12 2004 Received: with ECARTIS (v1.0.0; list csa); Fri, 15 Oct 2004 14:20:18 -0700 (PDT) Received: from andesite.usg.tufts.edu (andesite.usg.tufts.edu [130.64.1.202]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9FLKCme002546 for ; Fri, 15 Oct 2004 14:20:12 -0700 Received: from flux.usg.tufts.edu ([130.64.100.43]) by andesite.usg.tufts.edu with esmtp (Exim 4.20) id 1CIZUU-0007K2-jO; Fri, 15 Oct 2004 17:19:58 -0400 Date: Fri, 15 Oct 2004 17:19:58 -0400 From: Andrew Fant To: Jay Lan cc: Robin Holt , csa@oss.sgi.com Subject: Re: Two Simple Questions Message-ID: <134730000.1097875198@flux.usg.tufts.edu> In-Reply-To: <41703C56.3040805@sgi.com> References: <31770000.1097851964@flux.usg.tufts.edu> <20041015152049.GC873@lnx-holt.americas.sgi.com> <41700B10.1070102@sgi.com> <20041015173837.GG873@lnx-holt.americas.sgi.com> <20041015174254.GH873@lnx-holt.americas.sgi.com> <111360000.1097871666@flux.usg.tufts.edu> <41703754.2040409@sgi.com> <123050000.1097873483@flux.usg.tufts.edu> <41703C56.3040805@sgi.com> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-archive-position: 61 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: andrew.fant@tufts.edu Precedence: bulk X-list: csa --On Friday, October 15, 2004 14:08:38 -0700 Jay Lan wrote: > Yes, you were right. But then 'rmmod csa' cleans it up. Just > a workaround until the real fix is released. > > I will try to release the fixes later next week. It needs fixes > to both the rpm and kernel patch. > > - jay Ah, ok. Does this mean that CSA has to be used as a module and will break in other strange ways if I have it compiled into the kernel instead? Andy From jlan@engr.sgi.com Thu Oct 21 18:44:35 2004 Received: with ECARTIS (v1.0.0; list csa); Thu, 21 Oct 2004 18:44:41 -0700 (PDT) Received: from omx2.sgi.com (omx2-ext.sgi.com [192.48.171.19]) by oss.sgi.com (8.13.0/8.13.0) with ESMTP id i9M1iZ8M001273 for ; Thu, 21 Oct 2004 18:44:35 -0700 Received: from nodin.corp.sgi.com (nodin.corp.sgi.com [192.26.51.193]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9M2w0KA021540 for ; Thu, 21 Oct 2004 19:58:00 -0700 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by nodin.corp.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9M1iJT340206821 for ; Thu, 21 Oct 2004 18:44:19 -0700 (PDT) Received: from aware.engr.sgi.com (aware.engr.sgi.com [163.154.6.184]) by cthulhu.engr.sgi.com (SGI-8.12.5/8.12.5) with ESMTP id i9M1hILG5320997 for ; Thu, 21 Oct 2004 18:43:18 -0700 (PDT) Received: from engr.sgi.com (aware.engr.sgi.com [127.0.0.1]) by aware.engr.sgi.com (8.12.8/8.12.8) with ESMTP id i9M1hOEq024452 for ; Thu, 21 Oct 2004 18:43:24 -0700 Message-ID: <417865BC.6050803@engr.sgi.com> Date: Thu, 21 Oct 2004 18:43:24 -0700 From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: en-us, en MIME-Version: 1.0 To: CSA-ML Subject: [Fwd: [Lse-tech] [PATCH 2.6.9 0/2] enhanced accounting data collection] Content-Type: multipart/mixed; boundary="------------050909080707060108070707" X-archive-position: 62 X-ecartis-version: Ecartis v1.0.0 Sender: csa-bounce@oss.sgi.com Errors-to: csa-bounce@oss.sgi.com X-original-sender: jlan@engr.sgi.com Precedence: bulk X-list: csa This is a multi-part message in MIME format. --------------050909080707060108070707 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Here i forward my posting to lkml and lse-tech on the pathes to "common accouanting data collection" layer. If you are interested, please send follow-up to lse-tech. This new set of patches are free of CSA specific code. We try hard to get community to accept these patches. Your feedback on the patches or importance of this "enhanced accounting" certainly will help our case. Thanks! - jay --------------050909080707060108070707 Content-Type: message/rfc822; name="[Lse-tech] [PATCH 2.6.9 0/2] enhanced accounting data collection" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="[Lse-tech] [PATCH 2.6.9 0/2] enhanced accounting data collection" Return-Path: Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by aware.engr.sgi.com (8.12.8/8.12.8) with ESMTP id i9M1RTEq024427 for ; Thu, 21 Oct 2004 18:27:29 -0700 Received: from internal-mail-relay.corp.sgi.com (internal-mail-relay.corp.sgi.com [198.149.32.51]) by cthulhu.engr.sgi.com (SGI-8.12.5/8.12.5) with ESMTP id i9M1RILI5319334 for ; Thu, 21 Oct 2004 18:27:22 -0700 (PDT) Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by internal-mail-relay.corp.sgi.com (8.12.9/8.12.10/SGI_generic_relay-1.2) with ESMTP id i9M1RElb144291741; Thu, 21 Oct 2004 18:27:14 -0700 (PDT) X-ASG-Debug-ID: 1098408413-28610-10-0 X-Barracuda-URL: http://cuda.sgi.com:80/cgi-bin/mark.cgi Received: from sc8-sf-list1.sourceforge.net (lists.sourceforge.net [66.35.250.206]) by cuda.sgi.com (Spam Firewall) with ESMTP id 17327D028DC3; Thu, 21 Oct 2004 20:26:53 -0500 (CDT) Received: from localhost ([127.0.0.1] helo=projects.sourceforge.net) by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30) id 1CKoBg-0008Gm-1O; Thu, 21 Oct 2004 18:25:48 -0700 Received: from sc8-sf-mx1-b.sourceforge.net ([10.3.1.11] helo=sc8-sf-mx1.sourceforge.net) by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30) id 1CKo5F-0006u7-8e for lse-tech@lists.sourceforge.net; Thu, 21 Oct 2004 18:19:09 -0700 Received: from omx2-ext.sgi.com ([192.48.171.19] helo=omx2.sgi.com) by sc8-sf-mx1.sourceforge.net with esmtp (Exim 4.41) id 1CKo5E-0003s4-Po for lse-tech@lists.sourceforge.net; Thu, 21 Oct 2004 18:19:09 -0700 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by omx2.sgi.com (8.12.11/8.12.9/linux-outbound_gateway-1.1) with ESMTP id i9M2W55Q018736; Thu, 21 Oct 2004 19:32:05 -0700 Received: from aware.engr.sgi.com (aware.engr.sgi.com [163.154.6.184]) by cthulhu.engr.sgi.com (SGI-8.12.5/8.12.5) with ESMTP id i9M1INLG5317848; Thu, 21 Oct 2004 18:18:24 -0700 (PDT) Received: from engr.sgi.com (aware.engr.sgi.com [127.0.0.1]) by aware.engr.sgi.com (8.12.8/8.12.8) with ESMTP id i9M1IREq024396; Thu, 21 Oct 2004 18:18:27 -0700 Message-ID: <41785FE3.806@engr.sgi.com> From: Jay Lan User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 X-Accept-Language: en-us, en MIME-Version: 1.0 To: lse-tech CC: Andrew Morton , LKML , Guillaume Thouvenin Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Score: 0.0 (/) X-ASG-Orig-Subj: [Lse-tech] [PATCH 2.6.9 0/2] enhanced accounting data collection Subject: [Lse-tech] [PATCH 2.6.9 0/2] enhanced accounting data collection Sender: lse-tech-admin@lists.sourceforge.net Errors-To: lse-tech-admin@lists.sourceforge.net X-BeenThere: lse-tech@lists.sourceforge.net X-Mailman-Version: 2.0.9-sf.net Precedence: bulk List-Unsubscribe: , List-Id: List-Post: List-Help: List-Subscribe: , List-Archive: X-Original-Date: Thu, 21 Oct 2004 18:18:27 -0700 Date: Thu, 21 Oct 2004 18:18:27 -0700 X-Barracuda-Spam-Score: 0.00 X-Barracuda-Spam-Status: No, SCORE=0.00 using per-user scores of TAG_LEVEL=3.5 QUARANTINE_LEVEL=1000.0 KILL_LEVEL=1000.0 tests= X-Barracuda-Spam-Report: Code version 2.64, rules version 2.1.449 Rule breakdown below pts rule name description ---- ---------------------- ------------------------------------------- These two patches are the one we submitted to SuSE for Sles9 SP1. They are clean of CSA specific code. In earlier round of discussion, all partipants favored a common layer of accounting data collection. I believe these two patches are the super set that meets the needs of people who need enhanced BSD accounting. This patchset consists of two parts: acct_io and acct_mm, as we identified improved data collection in the area of IO and MM are useful to our customers. It is intended to offer common data collection method for various accounting packages including BSD accouting, ELSA, CSA, and any other acct packages that favor a common layer of data collection. 'acct_mm' defines a few macros that are no-op unless CONFIG_BSD_PROCESS_ACCT config flag is set on. Andrew, please consider including these two patches. Please let me know how i can help! Best Regards, --- Jay Lan - Linux System Software Silicon Graphics Inc., Mountain View, CA ------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT products in your business? Tell us what you think of them. Give us Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more http://productguide.itmanagersjournal.com/guidepromo.tmpl _______________________________________________ Lse-tech mailing list Lse-tech@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lse-tech --------------050909080707060108070707--