From owner-pcp@oss.sgi.com Mon Mar 5 10:42:41 2001 Received: by oss.sgi.com id ; Mon, 5 Mar 2001 10:42:32 -0800 Received: from mail.informatik.uni-ulm.de ([134.60.68.63]:5915 "EHLO mail.informatik.uni-ulm.de") by oss.sgi.com with ESMTP id ; Mon, 5 Mar 2001 10:42:26 -0800 Received: from [134.60.209.174] (helo=wsl10.informatik.uni-ulm.de) by mail.informatik.uni-ulm.de with esmtp (Exim 3.00 #1) id 14Zzvy-0006nj-00 for pcp@oss.sgi.com; Mon, 05 Mar 2001 19:42:14 +0100 Received: (from al1@localhost) by wsl10.informatik.uni-ulm.de (8.10.2/8.10.2/SuSE Linux 8.10.0-0.3) id f25IgDJ15555 for pcp@oss.sgi.com; Mon, 5 Mar 2001 19:42:13 +0100 From: Message-Id: <200103051842.f25IgDJ15555@wsl10.informatik.uni-ulm.de> Subject: pcp install dirs To: PCP Mails Date: Mon, 5 Mar 2001 19:42:13 +0100 (CET) X-Mailer: ELM [version 2.4ME+ PL60 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing hello, i want to build manually the pcp package. At this moment i have some problems related to specifying the install-dir-tree. Normally pcp is spreaded over the hole "unix" file system: /usr, /etc, /var, /usr/man, etc. I wish to install the hole package in only one directory. For example: /usr/local/pcp-2.1.10/{bin,share,lib,sbin,etc,include,var}. How can I do that (with minimal effort)? In the standard docs i didn't found any hints about that. Thanks for any hints, Adrian From owner-pcp@oss.sgi.com Mon Mar 5 14:35:23 2001 Received: by oss.sgi.com id ; Mon, 5 Mar 2001 14:35:03 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:12599 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Mon, 5 Mar 2001 14:34:40 -0800 Received: from rattle.melbourne.sgi.com (rattle.melbourne.sgi.com [134.14.55.145]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id OAA22939 for ; Mon, 5 Mar 2001 14:33:33 -0800 (PST) mail_from (kenmcd@melbourne.sgi.com) Received: from localhost (kenmcd@localhost) by rattle.melbourne.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id JAA04244; Tue, 6 Mar 2001 09:33:19 +1100 (AEDT) X-Authentication-Warning: rattle.melbourne.sgi.com: kenmcd owned process doing -bs Date: Tue, 6 Mar 2001 09:33:19 +1100 From: Ken McDonell Reply-To: To: cc: PCP Mails Subject: Re: pcp install dirs In-Reply-To: <200103051842.f25IgDJ15555@wsl10.informatik.uni-ulm.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing On Mon, 5 Mar 2001 al1@informatik.uni-ulm.de wrote: > hello, > > i want to build manually the pcp package. At this moment i have > some problems related to specifying the install-dir-tree. > Normally pcp is spreaded over the hole "unix" file system: > /usr, /etc, /var, /usr/man, etc. I wish to install the > hole package in only one directory. For example: > /usr/local/pcp-2.1.10/{bin,share,lib,sbin,etc,include,var}. > > How can I do that (with minimal effort)? > In the standard docs i didn't found any hints about that. I'd start with configure.in and make changes below this block dnl dnl Work out where to install stuff for this package dnl (and where to find stuff at run-time for add-on packages). dnl There are somethings the _have_ to go into / to work properly, e.g. the magic file for file(1), and whatever rc.d glue you use to get pmcd and pmie started. But most of the other stuff can be relocated based on the macros from configure.in that get translated into parameters in src/include/pcp.conf (which gets installed in /etc/pcp.conf and is used by /etc/pcp.env) ... all of the build-time and run-time path selections are based on the parameter settings in these files. From owner-pcp@oss.sgi.com Wed Mar 7 08:16:46 2001 Received: by oss.sgi.com id ; Wed, 7 Mar 2001 08:16:26 -0800 Received: from mail.informatik.uni-ulm.de ([134.60.68.63]:22808 "EHLO mail.informatik.uni-ulm.de") by oss.sgi.com with ESMTP id ; Wed, 7 Mar 2001 08:16:21 -0800 Received: from [134.60.209.174] (helo=wsl10.informatik.uni-ulm.de) by mail.informatik.uni-ulm.de with esmtp (Exim 3.00 #1) id 14agbq-0003zN-00; Wed, 07 Mar 2001 17:16:18 +0100 Received: (from al1@localhost) by wsl10.informatik.uni-ulm.de (8.10.2/8.10.2/SuSE Linux 8.10.0-0.3) id f27GGIa02347; Wed, 7 Mar 2001 17:16:18 +0100 From: Message-Id: <200103071616.f27GGIa02347@wsl10.informatik.uni-ulm.de> Subject: Re: pcp install dirs In-Reply-To: from Ken McDonell at "Mar 6, 2001 09:33:19 am" To: kenmcd@sgi.com Date: Wed, 7 Mar 2001 17:16:18 +0100 (CET) CC: PCP Mails X-Mailer: ELM [version 2.4ME+ PL60 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing Hello Ken, thank you for your hints. I could install pcp in the specified dirs. Adrian > On Mon, 5 Mar 2001 al1@informatik.uni-ulm.de wrote: > > > hello, > > > > i want to build manually the pcp package. At this moment i have > > some problems related to specifying the install-dir-tree. > > Normally pcp is spreaded over the hole "unix" file system: > > /usr, /etc, /var, /usr/man, etc. I wish to install the > > hole package in only one directory. For example: > > /usr/local/pcp-2.1.10/{bin,share,lib,sbin,etc,include,var}. > > > > How can I do that (with minimal effort)? > > In the standard docs i didn't found any hints about that. > > I'd start with > > configure.in > > and make changes below this block > > > dnl > dnl Work out where to install stuff for this package > dnl (and where to find stuff at run-time for add-on packages). > dnl > > There are somethings the _have_ to go into / to work properly, e.g. the > magic file for file(1), and whatever rc.d glue you use to get pmcd and > pmie started. > > But most of the other stuff can be relocated based on the macros from > configure.in that get translated into parameters in > src/include/pcp.conf (which gets installed in /etc/pcp.conf and is used > by /etc/pcp.env) ... all of the build-time and run-time path selections > are based on the parameter settings in these files. > > From owner-pcp@oss.sgi.com Wed Mar 7 10:13:36 2001 Received: by oss.sgi.com id ; Wed, 7 Mar 2001 10:13:17 -0800 Received: from mail.informatik.uni-ulm.de ([134.60.68.63]:7464 "EHLO mail.informatik.uni-ulm.de") by oss.sgi.com with ESMTP id ; Wed, 7 Mar 2001 10:12:59 -0800 Received: from [134.60.209.174] (helo=wsl10.informatik.uni-ulm.de) by mail.informatik.uni-ulm.de with esmtp (Exim 3.00 #1) id 14aiQi-0004R8-00 for pcp@oss.sgi.com; Wed, 07 Mar 2001 19:12:56 +0100 Received: (from al1@localhost) by wsl10.informatik.uni-ulm.de (8.10.2/8.10.2/SuSE Linux 8.10.0-0.3) id f27ICuA02857 for pcp@oss.sgi.com; Wed, 7 Mar 2001 19:12:56 +0100 Message-Id: <200103071812.f27ICuA02857@wsl10.informatik.uni-ulm.de> Subject: linux interface, pcp protocol, tutorials To: PCP Mails Date: Wed, 7 Mar 2001 19:12:56 +0100 (CET) From: adrian.lukas@informatik.uni-ulm.de Reply: adrian.lukas@informatik.uni-ulm.de X-Mailer: ELM [version 2.4ME+ PL60 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing hello everybody, i have the following questions: 1. Are any tutorials that describe some "pcp linux sessions"? In the "Admin's Guide" i read something about a "Tutorial". Is this tutorial available only for irix? After building and installing the package, i made some tests (pminfo). There were no metrics defined in any metric name space. It takes me some time to realize that the standard name spaces must be builded too. Unfortunately in the standard docs i didn't found any remarks about that. 2. In the "Admin's Guide" are some pcp tools described that are not present on linux distribution (at least on my box). Is that normal or i forget something? (example: pmchart) 3. What interface uses pcp to collect data on a linux host? Here i mean operating system specific data (linux metric name space), that refer to os activity (memory, processes, etc). Is /proc fs the "hole" linux interface? 4. Where can i find a specification of the pcp communication protocol? (here i mean "pcp" protocol, not TCP/IP). (I read something about a text and a binary protocol...) thanks in advance! adrian From owner-pcp@oss.sgi.com Thu Mar 8 23:36:24 2001 Received: by oss.sgi.com id ; Thu, 8 Mar 2001 23:36:15 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:7025 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Thu, 8 Mar 2001 23:35:58 -0800 Received: from cthulhu.engr.sgi.com (gate3-relay.engr.sgi.com [130.62.1.234]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id XAA11565 for ; Thu, 8 Mar 2001 23:34:51 -0800 (PST) mail_from (max@kuku.melbourne.sgi.com) Received: from kuku.melbourne.sgi.com (kuku.melbourne.sgi.com [134.14.55.163]) by cthulhu.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id XAA36644; Thu, 8 Mar 2001 23:34:39 -0800 (PST) Received: (from max@localhost) by kuku.melbourne.sgi.com (SGI-8.9.3/8.9.3) id SAA18937; Fri, 9 Mar 2001 18:34:35 +1100 (EST) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15016.34699.850984.926670@kuku.melbourne.sgi.com> Date: Fri, 9 Mar 2001 18:34:35 +1100 (EST) From: Max Matveev To: adrian.lukas@informatik.uni-ulm.de Cc: PCP Mails Subject: Re: linux interface, pcp protocol, tutorials In-Reply-To: <200103071812.f27ICuA02857@wsl10.informatik.uni-ulm.de> References: <200103071812.f27ICuA02857@wsl10.informatik.uni-ulm.de> X-Mailer: VM 6.72 under 21.1 (patch 10) "Capitol Reef" XEmacs Lucid Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing >>>>> "al" == adrian lukas writes: al> 1. Are any tutorials that describe some "pcp linux sessions"? al> In the "Admin's Guide" i read something about a "Tutorial". al> Is this tutorial available only for irix? At the moment - yes. It could appear on Linux one day, provided we will find time to check for Irix-isms and update it where necessary. al> After building and installing the package, i made some tests al> (pminfo). There were no metrics defined in any metric name al> space. It takes me some time to realize that the standard name al> spaces must be builded too. Unfortunately in the standard docs i al> didn't found any remarks about that. There is a bit of magic which is done for the user who installs RPMs - if file /var/pcp/pmns/.NeedRebuild exist, next time /etc/rc.d/init.d/pcp starts, it will rebuild the name space. Unfortunately, there is no install line for that file - we don't want it to be part of RPM because rpm(1) will complain after we remove it. In general, we don't expect Mrs User to do it - it's done automagically by pcp startup script and it's done by PMDAs' Install/Remove scripts, whence lack of documentation. BTW, there is also /var/pcp/config/magic/.NeedUpdate. al> 2. In the "Admin's Guide" are some pcp tools described that al> are not present on linux distribution (at least on my box). al> Is that normal or i forget something? (example: pmchart) It's normal: those tools are in the pcp-pro - another RPM which we ship as binary only. al> 3. What interface uses pcp to collect data on a linux host? al> Here i mean operating system specific data (linux metric name space), al> that refer to os activity (memory, processes, etc). al> Is /proc fs the "hole" linux interface? "hole" as in "whole"? Mostly yes. Some people refer to PCP as "distributed /proc". al> 4. Where can i find a specification of the pcp communication al> protocol? (here i mean "pcp" protocol, not TCP/IP). Do you mean over the wire protocol? I'm affraid there is no formal written specification, or at least I haven't seen one yet, - but sniffing around src/libpcp/src/pdu.c will uncover few interesting things. al> (I read something about a text and a binary protocol...) We normally use binary protocol, ie. values are sent as numbers and not as ascii strings. Ascii protocol is mostly for debugging. max From owner-pcp@oss.sgi.com Wed Mar 14 15:26:33 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 15:26:13 -0800 Received: from tomts5.bellnexxia.net ([209.226.175.25]:59292 "EHLO tomts5-srv.bellnexxia.net") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 15:26:12 -0800 Received: from avicenna ([64.230.196.36]) by tomts5-srv.bellnexxia.net (InterMail vM.4.01.03.16 201-229-121-116-20010115) with SMTP id <20010314232543.MSMF2129.tomts5-srv.bellnexxia.net@avicenna>; Wed, 14 Mar 2001 18:25:43 -0500 Content-Type: text/plain; charset="iso-8859-1" From: Anas Nashif Reply-To: nashif@suse.de To: Keith Owens Subject: Re: pcp build fails on ia64 Date: Wed, 14 Mar 2001 18:28:46 -0500 X-Mailer: KMail [version 1.2] Cc: pcp@oss.sgi.com References: <5147.982305263@kao2.melbourne.sgi.com> In-Reply-To: <5147.982305263@kao2.melbourne.sgi.com> Organization: SuSE MIME-Version: 1.0 Message-Id: <01031418284609.30623@avicenna> Content-Transfer-Encoding: 8bit Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing On February 16, 2001 01:34 am, Keith Owens wrote: > On Fri, 16 Feb 2001 05:43:19 +0100 (MET), > > nashif@suse.de wrote: > >gcc is > >gcc version 2.96-ia64-000717 snap 001117 > > > >gcc -Wall -L../../../src/libpcp/src -L../../../src/libpcp_pmda/src > > -shared -Wl,-soname,libpcp.so.2 -o libpcp.so.2 \ connect.o context.o > > desc.o err.o fetch.o freeresult.o help.o instance.o p_desc.o p_error.o > > p_fetch.o p_instance.o p_profile.o p_result.o p_text.o p_pmns.o p_creds.o > > pdu.o pdubuf.o pmns.o pool.o profile.o store.o units.o util.o ipc.o > > sortinst.o logmeta.o logportmap.o logutil.o tz.o interp.o checksum.o > > rtime.o tv.o mspec.o fetchlocal.o optfetch.o stuffvalue.o endian.o > > config.o access.o pmauth.o AF.o p_lcontrol.o p_lrequest.o p_lstatus.o > > logconnect.o logcontrol.o -ldl > > /usr/lib/gcc-lib/ia64-suse-linux/2.96-ia64-000717/../../../../ia64-suse-l > >inux/bin/ld: connect.o: @gprel relocation against dynamic symbol > > __pmConnectHostMethod > > This looks like a compiler bug. The code compiles fine on gcc > 2.9-ia64-000216-final, it fails to link when compiled with gcc > 2.96-ia64-000717 snap 001117. We are checking the latest gcc at the > moment to see if they have fixed this bug. I tried compiling with -fpic defined in src/include/builddefs and it seems to compile. >From our porting-howto: @gprel relocation against dynamic symbol ... This means that non-pic object files are linked into a shared library. The ia64 compiler _requires_ that all code in a shared library is compiled with -fpic. Anas From owner-pcp@oss.sgi.com Wed Mar 14 16:22:34 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 16:22:24 -0800 Received: from sgi.SGI.COM ([192.48.153.1]:20516 "EHLO sgi.com") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 16:22:11 -0800 Received: from larry.melbourne.sgi.com ([134.14.52.130]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via SMTP id QAA02183 for ; Wed, 14 Mar 2001 16:21:49 -0800 (PST) mail_from (markgw@sgi.com) Received: from sandpit.melbourne.sgi.com (sandpit.melbourne.sgi.com [134.14.55.132]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA23898; Thu, 15 Mar 2001 11:20:29 +1100 Date: Thu, 15 Mar 2001 11:20:29 +1100 (EST) From: Mark Goodwin X-Sender: markgw@sandpit.melbourne.sgi.com To: Anas Nashif cc: Keith Owens , pcp@oss.sgi.com Subject: Re: pcp build fails on ia64 In-Reply-To: <01031418284609.30623@avicenna> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pcp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pcp-outgoing On Wed, 14 Mar 2001, Anas Nashif wrote: > I tried compiling with -fpic defined in src/include/builddefs and it seems to > compile. > > >From our porting-howto: > > @gprel relocation against dynamic symbol ... > This means that non-pic object files are linked into a shared library. > The ia64 compiler _requires_ that all code in a shared library is compiled > with -fpic. We discovered the -fpic fix about a week ago and I've been meaning to dig out the original mail thread to let you know of the fix. But you beat me to it .. For PCP, the patch is to edit src/include/builddefs.in and change CFLAGS near line 70 (or so) to something like: CFLAGS += -fpic -fno-strict-aliasing $(LCFLAGS) $(WARN_OFF) -g -DPCP_DEBUG \ -DPCP_VERSION=\"$(PCP_VERSION)\" -I$(TOPDIR)/src/include Sorry I don't have the 2.1.x src handy at the moment, else I'd send you a proper diff/patch. BTW, the -fno-strict-aliasing is merely a precaution - we have no evidence it is actually needed, but pcp does some fancy pointer swizzling with it's pdu buffers, so it's in there just in case ... thanks -- Mark From owner-pcp@oss.sgi.com Sun Mar 25 17:08:05 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2Q185230371 for pcp-outgoing; Sun, 25 Mar 2001 17:08:05 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2Q184M30368 for ; Sun, 25 Mar 2001 17:08:04 -0800 Received: from larry.melbourne.sgi.com (larry.melbourne.sgi.com [134.14.52.130]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via SMTP id RAA19770 for ; Sun, 25 Mar 2001 17:06:50 -0800 (PST) mail_from (kenmcd@melbourne.sgi.com) From: kenmcd@melbourne.sgi.com Received: from [192.82.201.235] ([192.82.201.235]) by larry.melbourne.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id LAA13785; Mon, 26 Mar 2001 11:06:30 +1000 Date: Mon, 26 Mar 2001 11:12:45 +1000 (EST) Reply-To: kenmcd@melbourne.sgi.com To: adrian.lukas@informatik.uni-ulm.de cc: PCP Mails Subject: Re: linux interface, pcp protocol, tutorials In-Reply-To: <200103071812.f27ICuA02857@wsl10.informatik.uni-ulm.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pcp@oss.sgi.com Precedence: bulk Max answered much of this in a more timely fashion than I, so I'll just add a few snippets ... On Wed, 7 Mar 2001 adrian.lukas@informatik.uni-ulm.de wrote: > hello everybody, > > i have the following questions: > > 1. Are any tutorials that describe some "pcp linux sessions"? > In the "Admin's Guide" i read something about a "Tutorial". > Is this tutorial available only for irix? I've asked before (and got no response), seeking volunteers to help us transform the Tutorial from HTML documents written for IRIX users to something more OS neutral. It is not a big task, but some outside help would certainly be appreciated. > 2. In the "Admin's Guide" are some pcp tools described that > are not present on linux distribution (at least on my box). > Is that normal or i forget something? (example: pmchart) In general terms the GUI tools and some collector plugins from SGI are proprietary and intended to be used to manage the performance of SGI servers. pmchart and pmview are the most obvious applications in this group. > ... > 4. Where can i find a specification of the pcp communication > protocol? (here i mean "pcp" protocol, not TCP/IP). > (I read something about a text and a binary protocol...) Since the communication is encapsulated in either pmcd or libraries, for all the cases that are important most people (other than me!) can safely ignore the protocol specifications (and they are not written down anywhere). If you want to "see" the protocol in action, use -Dpdu on the command line to any PCP monitoring tool and you'll see the messages that are sent and received. Beyond that src/libpcp/pdu.c and src/libpcp/p_*.c are the places to look. We use the binary protocol in all places (except the news PMDA, which is a pretty pointless example, other than to reinforce why the ascii protocol was never a good idea!). With hindsight (and if it had even been an option at the time) investing in perl bindings to encapsulate the binary protocols would have been smarter.