From ciril at hcl.in Wed Feb 1 03:57:06 2006 From: ciril at hcl.in (Ciril Ignatious T) Date: Wed, 01 Feb 2006 09:27:06 +0530 Subject: [Linux-cluster] Linux Cluster In-Reply-To: <20060131124751.67730.qmail@web52307.mail.yahoo.com> References: <20060131124751.67730.qmail@web52307.mail.yahoo.com> Message-ID: <43E03192.6090204@hcl.in> Dear Suvankar Oracle RAC itself is a failover and loadbalancing cluster.So there is no need for a cluster suite for the HA functionality. You can directly install RHEL 3.0/4.0 in both machines and configure 10 RAC on top of OS. There is a very good documentation in the 10g RAC CD on how to set up the cluster. Also what do you mean by OS level cluster? Regards Ciril SUVANKAR MOITRA wrote: > dear rajesh, > > Then what about Oracle 10g rac? How oracle 10g rac to > be install ? I am using rhel as4 , redhat cluster > suite4 and oracle 10g rac. I want to build a cluster > in os level as well as database level. > > I am using follwing Hardware:---- > > 1> Hp dl380g4 x 2nos > 2> Hp Msa500 storage x 1 no > > what are the steps for the follwing installation.Pl > send me a proper document for the above. > > > > regards > > Suvankar Moitra > > > --- Rajesh singh wrote: > > >> Dear Suvankar, >> Its not quite clear as to what is your requirement. >> do you have a copy of RHE 3/4 and a copy of cluster >> suite >> if yes than install OS on either server, connect >> MSA500, install cluster >> suite on both machine and configure cluster. >> regards >> >> On 1/10/06, Suvankar Moitra >> wrote: >> >>> Hi, >>> >>> >>> I'm looking for an installing and configuring >>> >> procedure for RedHat Cluster >> >>> suite . >>> We have 2 servers HP Proliant DL380 G4 (RedHat >>> >> Advanced Server 4) attached >> >>> by HP MSA500 storage and we want to install and >>> >> configure Cluster suite . >> >>> Thanks. >>> >>> >>> Suvankar >>> >>> >>> -- >>> Linux-cluster mailing list >>> Linux-cluster at redhat.com >>> >>> > https://www.redhat.com/mailman/listinfo/linux-cluster > >>> -- >>> >> Linux-cluster mailing list >> Linux-cluster at redhat.com >> >> > https://www.redhat.com/mailman/listinfo/linux-cluster > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > > DISCLAIMER: > -------------------------------------------------------------------------------------------------------------- > > This e-mail contains confidential and/or privileged information. If you are not the intended recipient > (or have received this e-mail in error)please notify the sender immediately and destroy this e-mail. > Any unauthorized copying, disclosure, use or distribution of the material in this e-mail is strictly forbidden. > --------------------------------------------------------------------------------------------------------------- > > -- CIRIL IGNATIOUS T R & D ENGINEER HCL INFOSYSTEMS LTD PONDICHERRY PH:09894027005 -------------- next part -------------- An HTML attachment was scrubbed... URL: From suvankar_moitra at yahoo.com Wed Feb 1 06:04:21 2006 From: suvankar_moitra at yahoo.com (SUVANKAR MOITRA) Date: Tue, 31 Jan 2006 22:04:21 -0800 (PST) Subject: [Linux-cluster] using GNBD with other cluster filesystem In-Reply-To: <4c8edc920601302040n55d02bf3lceeb16b1528ead2a@mail.gmail.com> Message-ID: <20060201060421.79679.qmail@web52306.mail.yahoo.com> dear gourav, Today i am installing the above if iam sucessful i will send the doc to u. regards Suvankar --- gaurav wrote: > I had not exlored about using iSCSI with Oracle RAC. > i am now looking into that more seriously.(pl. pass > on any links if u have ) > > Yes, since this is only for testing purposes,so i do > not intend to use > multiple GNBD servers so i would not require to use > the -c (caching) > option > > regards > > gaurav > > > >hi > >If you use gnbd to export storage, the gnbd clients > should view this as a > >regular shared block device. You are definitely > able to put other filesystems > >on top of GNBD... with some caveats. GNBD is tied > pretty closely with the > >RHCS cluster manager. The only way you can export > devices uncached is with > >a cluster set up. If you don't need to export > uncached devices, you can use > >the -n (no cluster) option to avoid setting up a > RHCS cluster. > > >When devices are exported in cached mode, reads > will use the buffer cache on > >the GNBD server. This is a problem if you want to > use the exported block device > >directly on the server, or want to export the same > shared storage device from > >multiple gnbd servers. If you aren't doing either > of those, you should be able > >to use the -n option. > > >-Ben > > >Another alternative would be to use iscsi, with a > software iscsi target. > > > -- > my site:http://www.gnulinuxclub.org > my blog:http://linux4all.blogspot.com > my project:http://masand.sourceforge.net > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From libregeek at gmail.com Wed Feb 1 08:09:11 2006 From: libregeek at gmail.com (Manilal K M) Date: Wed, 1 Feb 2006 13:39:11 +0530 Subject: [Linux-cluster] using GNBD with other cluster filesystem In-Reply-To: <20060201060421.79679.qmail@web52306.mail.yahoo.com> References: <4c8edc920601302040n55d02bf3lceeb16b1528ead2a@mail.gmail.com> <20060201060421.79679.qmail@web52306.mail.yahoo.com> Message-ID: <2315046d0602010009s24ba1613n@mail.gmail.com> On 01/02/06, SUVANKAR MOITRA wrote: > dear gourav, > > Today i am installing the above if iam sucessful i > will send the doc to u. > Remember to send it to the mailing list From brodriguezb at fujitsu.es Wed Feb 1 11:38:42 2006 From: brodriguezb at fujitsu.es (=?UTF-8?B?QmFydG9sb23DqSBSb2Ryw61ndWV6?=) Date: Wed, 01 Feb 2006 12:38:42 +0100 Subject: [Linux-cluster] journal size gfs. Message-ID: <43E09DC2.6010306@fujitsu.es> Hello, i dont remember the journals size of my gfs partitions, how can i get it? id like add journal with the same size. I have used gfs_tool jindex but i dont see anything. Thanks and regards. Bart. -- ________________________________________ Bartolom? Rodr?guez Bordallo Departamento de Explotaci?n de Servicios FUJITSU ESPA?A SERVICES, S.A.U. Camino Cerro de los Gamos, 1 28224 Pozuelo de Alarc?n, Madrid Tel.: 902 11 40 10 Mail: brodriguezb at fujitsu.es ________________________________________ La informaci?n contenida en este e-mail es confidencial y va dirigida ?nicamente al receptor que aparece como destinatario. Si ha recibido este e-mail por error, por favor, notif?quenoslo inmediatamente y b?rrelo de su sistema. Por favor, en tal caso, no lo copie ni lo use para ning?n prop?sito, ni revele sus contenidos a ninguna persona ni lo almacene ni copie esta informaci?n en ning?n medio. From depeecmr at yahoo.com Wed Feb 1 11:43:21 2006 From: depeecmr at yahoo.com (Daniel EPEE LEA) Date: Wed, 1 Feb 2006 03:43:21 -0800 (PST) Subject: [Linux-cluster] GFS over WAN link implementation In-Reply-To: <43E09DC2.6010306@fujitsu.es> Message-ID: <20060201114321.10388.qmail@web30207.mail.mud.yahoo.com> Everyone, I have this setup (please check attached picture). WAN link has dedicated 2048 Kbps (with average 14 miliseconds between nodes ) LB 1, 2, 3 Servers are Load Balancers Node 1, 2, 3, 4 are Redhat Enterprise linux ES v4 + GFS MSA1000 storage has 30 Gig of Data right now, and can grow beyond 100Gigs. Can someone help me out to figure the best GFS configuration to have: 1- Mirrored storage over the Wan Link ? 2- What options can be setup for asynchroneous replication of the storrage if the WAN link is becomes too thin for synchronous replication ? 3- Nodes 3 & 4 fail over Nodes 1 & 2, supposing that Node 1 and 2 are unavailable ? 4- Does GFS over DRBD work at this time ? and have someone implement it ? Thanks for your help, Dan __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From gforte at leopard.us.udel.edu Wed Feb 1 14:49:40 2006 From: gforte at leopard.us.udel.edu (Greg Forte) Date: Wed, 01 Feb 2006 09:49:40 -0500 Subject: [Linux-cluster] verbose clurgmgrd output? In-Reply-To: <1138744624.4371.370.camel@ayanami.boston.redhat.com> References: <43DE9133.6000409@leopard.us.udel.edu> <1138744624.4371.370.camel@ayanami.boston.redhat.com> Message-ID: <43E0CA84.9020806@leopard.us.udel.edu> Lon Hohberger wrote: > On Mon, 2006-01-30 at 17:20 -0500, Greg Forte wrote: >> Is there any way to get clurgmgrd to output more verbosely, esp. >> whatever messages there might have been from scripts it's managing? >> I've got several cluster services that don't behave properly, in one >> case if I run the script manually (i.e. sudo /etc/init.d/scriptname), it >> works fine, but when clurgmgrd trys it says it "returned 1 (generic >> error)". Another appears to run perfectly fine (no error output from >> clurgmgrd), but as soon as it's up the manager stops it again, then >> repeats the cycle. Again, running it manually works fine. These are >> homespun scripts, so it's quite possible I'm missing something basic, >> but I can't figure out what (obviously ;-) > > Hi Greg, > > Try adding log_level="7" to the tag if you're using current > STABLE/RHEL4 branches (don't forget to update the configuration > version). Note that older versions of the resource scripts don't have > much in the way of logging. Hmmm ... tried this, but no apparent change. This would log into /var/log/messages, correct? I am running the latest RHEL4 packages. > Also, you can run "clurgmgrd -d", but that requires a restart. same result. > As an alternative -- you can run fun tests on the service manually: > > clusvcadm -d (service must be disabled!) > > rg_test test /etc/cluster/cluster.conf start service > rg_test test /etc/cluster/cluster.conf status service > rg_test test /etc/cluster/cluster.conf stop service rg_test segfaults immediately, no matter what I try. ;-) > (p.s. Say 'Hi' to mikeyp and monogoose for me if you see them. ^.^) I know mikeyp, but no clue who monogoose is ... -g Greg Forte gforte at udel.edu IT - User Services University of Delaware 302-831-1982 Newark, DE From gforte at leopard.us.udel.edu Wed Feb 1 14:57:05 2006 From: gforte at leopard.us.udel.edu (Greg Forte) Date: Wed, 01 Feb 2006 09:57:05 -0500 Subject: [Linux-cluster] verbose clurgmgrd output? In-Reply-To: <43E0CA84.9020806@leopard.us.udel.edu> References: <43DE9133.6000409@leopard.us.udel.edu> <1138744624.4371.370.camel@ayanami.boston.redhat.com> <43E0CA84.9020806@leopard.us.udel.edu> Message-ID: <43E0CC41.3090502@leopard.us.udel.edu> Never mind, I'm an idiot - none of my scripts support the "status" command, so of course the resource manager is going to fail them ... ugh. In my defense, though, I haven't found any documentation describing minimum requirements for a CS-compatible init script. Does any exist? Still be interesting to know what's up with rg_test, too ... -g Greg Forte wrote: > Lon Hohberger wrote: >> On Mon, 2006-01-30 at 17:20 -0500, Greg Forte wrote: >>> Is there any way to get clurgmgrd to output more verbosely, esp. >>> whatever messages there might have been from scripts it's managing? >>> I've got several cluster services that don't behave properly, in one >>> case if I run the script manually (i.e. sudo /etc/init.d/scriptname), >>> it works fine, but when clurgmgrd trys it says it "returned 1 >>> (generic error)". Another appears to run perfectly fine (no error >>> output from clurgmgrd), but as soon as it's up the manager stops it >>> again, then repeats the cycle. Again, running it manually works >>> fine. These are homespun scripts, so it's quite possible I'm missing >>> something basic, but I can't figure out what (obviously ;-) >> >> Hi Greg, >> >> Try adding log_level="7" to the tag if you're using current >> STABLE/RHEL4 branches (don't forget to update the configuration >> version). Note that older versions of the resource scripts don't have >> much in the way of logging. > > Hmmm ... tried this, but no apparent change. This would log into > /var/log/messages, correct? I am running the latest RHEL4 packages. > >> Also, you can run "clurgmgrd -d", but that requires a restart. > > same result. > >> As an alternative -- you can run fun tests on the service manually: >> >> clusvcadm -d (service must be disabled!) >> >> rg_test test /etc/cluster/cluster.conf start service >> rg_test test /etc/cluster/cluster.conf status service >> rg_test test /etc/cluster/cluster.conf stop service > > rg_test segfaults immediately, no matter what I try. ;-) > >> (p.s. Say 'Hi' to mikeyp and monogoose for me if you see them. ^.^) > > I know mikeyp, but no clue who monogoose is ... > > -g > > Greg Forte > gforte at udel.edu > IT - User Services > University of Delaware > 302-831-1982 > Newark, DE > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -- Greg Forte gforte at udel.edu IT - User Services University of Delaware 302-831-1982 Newark, DE From Alain.Moulle at bull.net Wed Feb 1 16:03:13 2006 From: Alain.Moulle at bull.net (Alain Moulle) Date: Wed, 01 Feb 2006 17:03:13 +0100 Subject: [Linux-cluster] CS4/ Heart-beat configuration (contd.) Message-ID: <43E0DBC1.7070504@bull.net> Alain Moulle wrote: >> Hi >> >> Is there a way to force the CS4 to use >> another interface for Heart-Beat than >> the one linked to the hostname ? >> And if so, how to ? >cman_tool join -n >see cman_tool -h (or the man page) for a list of options. >-- patrick Hi OK I have given this a try, and this does not work : in fact, it returns : not know in cluster.conf or something likewise. Knowing that, I have tried another thing : set a name on eth1 in /etc/hosts such as : 10.0.0.2 nodehb #for heart beat where is eth0 remains : 10.0.0.1 node (which also is the hostname) Then in cluster.conf, I have changed all "name" values from "node" to "nodehb" And this seems to work : I mean the Heart-beat frames are effectively running through eth1 now and I have tried to failover by doing ifdown eth1 on the first node and the failover has been successful. So it seems that the Heart-Beat is not linked to the IF linked to the hostname, but really linked only to the names we put in cluster.conf. Could someone confirm that point ? And eventually tell me which eventual drawback I could face to adopt this configuration ? Thanks a lot Alain Moull? From pcaulfie at redhat.com Wed Feb 1 17:20:01 2006 From: pcaulfie at redhat.com (Patrick Caulfield) Date: Wed, 01 Feb 2006 17:20:01 +0000 Subject: [Linux-cluster] CS4/ Heart-beat configuration (contd.) In-Reply-To: <43E0DBC1.7070504@bull.net> References: <43E0DBC1.7070504@bull.net> Message-ID: <43E0EDC1.20400@redhat.com> Alain Moulle wrote: > Alain Moulle wrote: > >>> Hi >>> >>> Is there a way to force the CS4 to use >>> another interface for Heart-Beat than >>> the one linked to the hostname ? >>> And if so, how to ? > > >> cman_tool join -n >> see cman_tool -h (or the man page) for a list of options. >> -- patrick > > Hi > OK I have given this a try, and this does not work : > in fact, it returns : not know in cluster.conf > or something likewise. > Knowing that, I have tried another thing : > set a name on eth1 in /etc/hosts such as : > 10.0.0.2 nodehb #for heart beat > where is eth0 remains : > 10.0.0.1 node (which also is the hostname) > Then in cluster.conf, I have changed all "name" values > from "node" to "nodehb" > And this seems to work : I mean the Heart-beat > frames are effectively running through eth1 now > and I have tried to failover by doing ifdown eth1 on > the first node and the failover has been successful. > > So it seems that the Heart-Beat is not linked > to the IF linked to the hostname, but really linked > only to the names we put in cluster.conf. It's related to both, as you found.... > Could someone confirm that point ? > And eventually tell me which eventual drawback I could > face to adopt this configuration ? > No, that's almost exactly what you're supposed to do :) -- patrick From Leonardo.Mello at planejamento.gov.br Wed Feb 1 17:45:41 2006 From: Leonardo.Mello at planejamento.gov.br (Leonardo Rodrigues de Mello) Date: Wed, 1 Feb 2006 15:45:41 -0200 Subject: [Linux-cluster] GFS performance analysis Message-ID: <1DDCE5B29CB5BC42BC2BFC39E3F1C8A3255AC4@corp-bsa-mp01.planejamento.gov.br> We are doing some research here in the brazilian government in distributed mass storage. one of the scenarios that will be study involve gfs, and we will test gfs performance localy, exported to another machine via gnbd, iscsi and enbd. we found some weird results, where my disks localy can reach over 300MB/sec using ext3, and the same test using gfs reach only 38MB/sec. There is one graphic about this analysis in the page, and results for dbench and iozone. In this moment we are doing benchmarks using bonnie++ to certify that this weird performance arent relate to the tools we had used. Other consideration is that turning off hyperthread improved a litle (6-10MB/s) the performance, i believe this is a bug, or one problem in design of gfs. is there a way to improve the gfs performance ? or this poor performance is all that i can get? Leonardo Mello -------------- next part -------------- An HTML attachment was scrubbed... URL: From Leonardo.Mello at planejamento.gov.br Wed Feb 1 17:50:17 2006 From: Leonardo.Mello at planejamento.gov.br (Leonardo Rodrigues de Mello) Date: Wed, 1 Feb 2006 15:50:17 -0200 Subject: [Linux-cluster] RE: GFS performance analysis Message-ID: <1DDCE5B29CB5BC42BC2BFC39E3F1C8A3255AC5@corp-bsa-mp01.planejamento.gov.br> Sorry i forgot the url: http://guialivre.governoeletronico.gov.br/mediawiki/index.php/TestesGFS ---------------------------------------------------------------------------- We are doing some research here in the brazilian government in distributed mass storage. one of the scenarios that will be study involve gfs, and we will test gfs performance localy, exported to another machine via gnbd, iscsi and enbd. we found some weird results, where my disks localy can reach over 300MB/sec using ext3, and the same test using gfs reach only 38MB/sec. There is one graphic about this analysis in the page, and results for dbench and iozone. In this moment we are doing benchmarks using bonnie++ to certify that this weird performance arent relate to the tools we had used. Other consideration is that turning off hyperthread improved a litle (6-10MB/s) the performance, i believe this is a bug, or one problem in design of gfs. is there a way to improve the gfs performance ? or this poor performance is all that i can get? Leonardo Mello -------------- next part -------------- An HTML attachment was scrubbed... URL: From lhh at redhat.com Wed Feb 1 18:09:26 2006 From: lhh at redhat.com (Lon Hohberger) Date: Wed, 01 Feb 2006 13:09:26 -0500 Subject: [Linux-cluster] verbose clurgmgrd output? In-Reply-To: <43E0CC41.3090502@leopard.us.udel.edu> References: <43DE9133.6000409@leopard.us.udel.edu> <1138744624.4371.370.camel@ayanami.boston.redhat.com> <43E0CA84.9020806@leopard.us.udel.edu> <43E0CC41.3090502@leopard.us.udel.edu> Message-ID: <1138817366.4371.380.camel@ayanami.boston.redhat.com> On Wed, 2006-02-01 at 09:57 -0500, Greg Forte wrote: > Never mind, I'm an idiot - none of my scripts support the "status" > command, so of course the resource manager is going to fail them ... ugh. > > In my defense, though, I haven't found any documentation describing > minimum requirements for a CS-compatible init script. Does any exist? Pretty sure any script correctly implementing the LSB-spec for init scripts should work. Check here: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html -- Lon From lhh at redhat.com Wed Feb 1 18:29:38 2006 From: lhh at redhat.com (Lon Hohberger) Date: Wed, 01 Feb 2006 13:29:38 -0500 Subject: [Linux-cluster] verbose clurgmgrd output? In-Reply-To: <43E0CA84.9020806@leopard.us.udel.edu> References: <43DE9133.6000409@leopard.us.udel.edu> <1138744624.4371.370.camel@ayanami.boston.redhat.com> <43E0CA84.9020806@leopard.us.udel.edu> Message-ID: <1138818578.4371.390.camel@ayanami.boston.redhat.com> On Wed, 2006-02-01 at 09:49 -0500, Greg Forte wrote: > Hmmm ... tried this, but no apparent change. This would log into > /var/log/messages, correct? I am running the latest RHEL4 packages. Yes, and they're not fixed in the current RHEL4 packages, but it will be in the next update. > > Also, you can run "clurgmgrd -d", but that requires a restart. > > same result. /var/log/messages filters out debug info. Keep using '-d' for now and add: daemon.* /var/log/daemon-log to /etc/syslog.conf. This should trap all messages from rgmanager, including debug info. > > > As an alternative -- you can run fun tests on the service manually: > > > > clusvcadm -d (service must be disabled!) > > > > rg_test test /etc/cluster/cluster.conf start service > > rg_test test /etc/cluster/cluster.conf status service > > rg_test test /etc/cluster/cluster.conf stop service > > rg_test segfaults immediately, no matter what I try. ;-) ... x86_64 ? See bz #177340. It's fixed in current CVS in all branches (or should be), and will also be fixed in the next update of the RHEL4 packages. > > (p.s. Say 'Hi' to mikeyp and monogoose for me if you see them. ^.^) > > I know mikeyp, but no clue who monogoose is ... One of mikeyp's minions. -- Lon From gforte at leopard.us.udel.edu Wed Feb 1 18:36:13 2006 From: gforte at leopard.us.udel.edu (Greg Forte) Date: Wed, 01 Feb 2006 13:36:13 -0500 Subject: [Linux-cluster] verbose clurgmgrd output? In-Reply-To: <1138817366.4371.380.camel@ayanami.boston.redhat.com> References: <43DE9133.6000409@leopard.us.udel.edu> <1138744624.4371.370.camel@ayanami.boston.redhat.com> <43E0CA84.9020806@leopard.us.udel.edu> <43E0CC41.3090502@leopard.us.udel.edu> <1138817366.4371.380.camel@ayanami.boston.redhat.com> Message-ID: <43E0FF9D.3010704@leopard.us.udel.edu> Cool, thanks. Heh, here's an interesting quote: "In addition to straightforward success, the following situations are also to be considered successful: ... running stop on a service already stopped or not running ..." This is exactly what I was complaining/asking about a few days ago - the "stock" httpd init script in RHEL4 returns non-zero if you try to stop it when it's not running. This in turn causes clusvcadm -d to fail if you try to disable it when it's not running, so you have to manually start it "non-clusterfied" before you can disable it in the cluster! I haven't done a survey, but I'm guessing other scripts behave similarly. But the spec is self-contradictory, because further down it defines an error status code for "program is not running" that it says it to be used for any action other than 'status'! So ... is this a bug in the script or a "bug" in the document? -g Lon Hohberger wrote: > On Wed, 2006-02-01 at 09:57 -0500, Greg Forte wrote: >> Never mind, I'm an idiot - none of my scripts support the "status" >> command, so of course the resource manager is going to fail them ... ugh. >> >> In my defense, though, I haven't found any documentation describing >> minimum requirements for a CS-compatible init script. Does any exist? > > Pretty sure any script correctly implementing the LSB-spec for init > scripts should work. Check here: > > http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html > > -- Lon > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -- Greg Forte gforte at udel.edu IT - User Services University of Delaware 302-831-1982 Newark, DE From aconrad.tlv at magic.fr Wed Feb 1 19:20:00 2006 From: aconrad.tlv at magic.fr (Alexandre CONRAD) Date: Wed, 01 Feb 2006 20:20:00 +0100 Subject: [Linux-cluster] distributed database on multiple servers/disks Message-ID: <43E109E0.5090402@magic.fr> Hello, I'm thinking about setting up a MySQL database clustered (but not like MySQL Cluster which kinda just does mirroring AFAIK). I want it to be partionned on multiple disks/servers. Here is my apporch: +-------------------+ | Application (PHP) | +-------------------+ ^ | | v +--------------+ | MySQL Server | +--------------+ ^ | | v +--------------- logical disk ----------------+ | ^ | | | | | +----LAN-----+----LAN-----+ | | | | | | | v v v | | +----------+ +----------+ +----------+ | | | physical | | physical | | physical | | | | server 1 | | server 2 | | server 3 | | | +----------+ +----------+ +----------+ | | | +---------------------------------------------+ The idea is to split the data on 3 physical servers when data is written. So each server will have 1/3 of the data on their disk. It would work as a RAID 5 system. In case of full disk space, you would add a new physical server (node) and the data would be redistributed to all servers, each getting 1/4 of the data. In case of server failure, because of distributed parity (RAID5), it could keep going. Or alternativly think about a RAID0+1 configuration. Do you have any idea if this could be achived using LVM, RH Cluster Suite or any other clustering / partitioning system ? Regards, -- Alexandre CONRAD - TLV Research & Development tel : +33 1 30 80 55 05 fax : +33 1 30 80 55 06 6, rue de la plaine 78860 - SAINT NOM LA BRETECHE FRANCE From gkapitany at rogers.com Wed Feb 1 19:57:56 2006 From: gkapitany at rogers.com (GABRIEL KAPITANY) Date: Wed, 1 Feb 2006 14:57:56 -0500 (EST) Subject: [Linux-cluster] distributed database on multiple servers/disks In-Reply-To: <43E109E0.5090402@magic.fr> Message-ID: <20060201195756.10877.qmail@web88106.mail.re2.yahoo.com> Hi, another approach, not clustering, which might give you a distributed database: https://forge.continuent.org/projects/sequoia/ Gabriel --- Alexandre CONRAD wrote: > Hello, > > I'm thinking about setting up a MySQL database > clustered (but not like > MySQL Cluster which kinda just does mirroring > AFAIK). > > I want it to be partionned on multiple > disks/servers. > > Here is my apporch: > > +-------------------+ > | Application (PHP) | > +-------------------+ > ^ > | > | > v > +--------------+ > | MySQL Server | > +--------------+ > ^ > | > | > v > +--------------- logical disk ----------------+ > | ^ | > | | | > | +----LAN-----+----LAN-----+ | > | | | | | > | v v v | > | +----------+ +----------+ +----------+ | > | | physical | | physical | | physical | | > | | server 1 | | server 2 | | server 3 | | > | +----------+ +----------+ +----------+ | > | | > +---------------------------------------------+ > > The idea is to split the data on 3 physical servers > when data is > written. So each server will have 1/3 of the data on > their disk. It > would work as a RAID 5 system. > > In case of full disk space, you would add a new > physical server (node) > and the data would be redistributed to all servers, > each getting 1/4 of > the data. > > In case of server failure, because of distributed > parity (RAID5), it > could keep going. > > Or alternativly think about a RAID0+1 configuration. > > Do you have any idea if this could be achived using > LVM, RH Cluster > Suite or any other clustering / partitioning system > ? > > Regards, > -- > Alexandre CONRAD - TLV > Research & Development > tel : +33 1 30 80 55 05 > fax : +33 1 30 80 55 06 > 6, rue de la plaine > 78860 - SAINT NOM LA BRETECHE > FRANCE > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > From Anthony.Assi at irisa.fr Thu Feb 2 13:14:37 2006 From: Anthony.Assi at irisa.fr (Anthony Assi) Date: Thu, 02 Feb 2006 14:14:37 +0100 Subject: [Linux-cluster] RE: GFS performance analysis In-Reply-To: <1DDCE5B29CB5BC42BC2BFC39E3F1C8A3255AC5@corp-bsa-mp01.planejamento.gov.br> References: <1DDCE5B29CB5BC42BC2BFC39E3F1C8A3255AC5@corp-bsa-mp01.planejamento.gov.br> Message-ID: <43E205BD.2070301@irisa.fr> Leonardo Rodrigues de Mello wrote: > Sorry i forgot the url: > > http://guialivre.governoeletronico.gov.br/mediawiki/index.php/TestesGFS > > ---------------------------------------------------------------------------- > We are doing some research here in the brazilian government in > distributed mass storage. > > one of the scenarios that will be study involve gfs, and we will test > gfs performance localy, exported to another machine via gnbd, iscsi > and enbd. > > we found some weird results, where my disks localy can reach over > 300MB/sec using ext3, and the same test using gfs reach only 38MB/sec. > > There is one graphic about this analysis in the page, and results for > dbench and iozone. > > In this moment we are doing benchmarks using bonnie++ to certify that > this weird performance arent relate to the tools we had used. > > Other consideration is that turning off hyperthread improved a litle > (6-10MB/s) the performance, i believe this is a bug, or one problem in > design of gfs. > > is there a way to improve the gfs performance ? or this poor > performance is all that i can get? > > Leonardo Mello > >------------------------------------------------------------------------ > >-- >Linux-cluster mailing list >Linux-cluster at redhat.com >https://www.redhat.com/mailman/listinfo/linux-cluster > Dear Leonardo, We are facing the same problem on our cluster, We are actually using the GULM lock system, and are hoping that by switching to the DLM locking system, we will resolve this problem. Wich locking method are you using for GFS? I will be gratefull, if you could keep me informed of your upcomming performance/solutions. Regards, -- Anthony Assi DBA/System Administrator Bio-Informatics Platform, Symbiose Team IRISA - INRIA, Rennes, France Tel: +33 2 99 84 71 58 http://genouest.org From sp at linworx.cz Thu Feb 2 13:58:17 2006 From: sp at linworx.cz (=?ISO-8859-2?Q?Stanislav_Pol=E1=B9ek?=) Date: Thu, 02 Feb 2006 14:58:17 +0100 Subject: [Linux-cluster] two nodes -> tree nodes Message-ID: <43E20FF9.4060406@linworx.cz> Hi everybody. I would like to add third node to the running cluster without restart of the cluster services. But when I try to join the cluster with the third node, the other two refuse the requests with the message "CMAN: join request from node requested, exceeds two node limit". My cluster.conf does not explicitly state the two_node parameter, but cman seems to run in that mode. What would be the right way to add the third node without restart of the cluster services on the other two? Thank you very much Stanislav -- #--#--#--#--#--#--#--#--#--#-- Stanislav Polasek RHCE #807302906006864 LinWorx s.r.o. sp at linworx.cz / www.linworx.cz #--#--#--#--#--#--#--#--#--#-- From leonardo.mello at planejamento.gov.br Thu Feb 2 14:37:11 2006 From: leonardo.mello at planejamento.gov.br (Leonardo Rodrigues de Mello) Date: Thu, 02 Feb 2006 12:37:11 -0200 Subject: [Linux-cluster] RE: GFS performance analysis In-Reply-To: <43E205BD.2070301@irisa.fr> References: <1DDCE5B29CB5BC42BC2BFC39E3F1C8A3255AC5@corp-bsa-mp01.planejamen to.gov.br> <43E205BD.2070301@irisa.fr> Message-ID: <1138891031.13467.10.camel@mp> Dear Anthony, The tests that we report in the last message, were done localy in the machine (without gnbd), and without lock (just one machine) using module lock_nolock. We believe the performance problem we faced is about gfs structure, not lock, but now, we are updating the webpage and results with GNBD and GFS will be available soon, may be today. just to remember the page: http://guialivre.governoeletronico.gov.br/seminario/index.php/TestesGFS Our objective here is to implement one Distributed Raid in a network enviroment using gfs as the top filesystem. I had done that with success about two years ago using ext3 at top filesystem. It was one distributed raid in a network enviroment of 1.5 TB distributed in 4 machine servers and 1 client mounting that raid. Lets see how gfs performs, in the next tests cya -- .''`. Leonardo Rodrigues de Mello : :' : Coordenador de Projetos em Cluster e Grid `. `' DSI/SLTI/MP `- 55 61 3313 1329 > Dear Leonardo, > > We are facing the same problem on our cluster, > We are actually using the GULM lock system, and are hoping that by > switching to the DLM locking system, we will resolve this problem. > Wich locking method are you using for GFS? > > I will be gratefull, if you could keep me informed of your upcomming > performance/solutions. > > Regards, > -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: From Anthony.Assi at irisa.fr Thu Feb 2 15:36:20 2006 From: Anthony.Assi at irisa.fr (Anthony Assi) Date: Thu, 02 Feb 2006 16:36:20 +0100 Subject: [Linux-cluster] /etc/cluster/cluster.conf Message-ID: <43E226F4.7030001@irisa.fr> I am debugging a Performance problem we are having with our Cluster, We have a 32 nodes cluster, using a GFS shared Volume, I would like to know whether this is a good configuration of the "cluster.conf" File: [root at genocluster-data symbiose]# more /etc/cluster/cluster.conf .............................................................................Copy/Paste for : genouest1 till genouest31............................................................................................. -- Anthony Assi DBA/System Administrator INRIA: French National Institute for Research in Computer Science and Control IRISA, Rennes, France Tel: +33 2 99 84 71 58 http://www.irisa.fr/symbiose/index-eng.htm From pcaulfie at redhat.com Thu Feb 2 16:15:14 2006 From: pcaulfie at redhat.com (Patrick Caulfield) Date: Thu, 02 Feb 2006 16:15:14 +0000 Subject: [Linux-cluster] two nodes -> tree nodes In-Reply-To: <43E20FF9.4060406@linworx.cz> References: <43E20FF9.4060406@linworx.cz> Message-ID: <43E23012.5010405@redhat.com> Stanislav Pol??ek wrote: > Hi everybody. I would like to add third node to the running cluster > without restart of the cluster services. But when I try to join the > cluster with the third node, the other two refuse the requests with the > message "CMAN: join request from node requested, exceeds two node > limit". My cluster.conf does not explicitly state the two_node > parameter, but cman seems to run in that mode. What would be the right > way to add the third node without restart of the cluster services on the > other two? > two_node="1" Must have been in the cluster.conf file when the cluster was formed - that's the only way the flag inside CMAN can get set. Once you have a 2 node cluster the only way to add a node is to remove all nodes from the cluster and join them again without two_node set. >From 3 node upwards you can add nodes as you please - it's just the transition from 2 to 3 that needs a restart because 2 nodes is a special case. -- patrick From brilong at cisco.com Thu Feb 2 16:40:34 2006 From: brilong at cisco.com (Brian Long) Date: Thu, 02 Feb 2006 11:40:34 -0500 Subject: [Linux-cluster] Errata Mailing List? Message-ID: <1138898434.4441.14.camel@brilong-lnx> Hello, I just joined the list and was wondering if there is a mailman list similar to enterprise-watch-list that gives errata notices for GFS and RHCS. Thanks. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From epeelea at gmail.com Thu Feb 2 21:03:44 2006 From: epeelea at gmail.com (Daniel EPEE LEA) Date: Thu, 2 Feb 2006 22:03:44 +0100 Subject: [Linux-cluster] Errata Mailing List? In-Reply-To: <1138898434.4441.14.camel@brilong-lnx> References: <1138898434.4441.14.camel@brilong-lnx> Message-ID: Hi Brian, I do not know about a list yet, but i know that if you have GFS licenses or RHEL Entitlements you will receive daily errata notices through your RHN account. That is how I get them. Hope this helps Daniel On 2/2/06, Brian Long wrote: > Hello, > > I just joined the list and was wondering if there is a mailman list > similar to enterprise-watch-list that gives errata notices for GFS and > RHCS. > > Thanks. > > /Brian/ > -- > Brian Long | | | > IT Data Center Systems | .|||. .|||. > Cisco Linux Developer | ..:|||||||:...:|||||||:.. > Phone: (919) 392-7363 | C i s c o S y s t e m s > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > From gforte at leopard.us.udel.edu Thu Feb 2 23:23:57 2006 From: gforte at leopard.us.udel.edu (Greg Forte) Date: Thu, 02 Feb 2006 18:23:57 -0500 Subject: [Linux-cluster] dependencies between services Message-ID: <43E2948D.3060108@leopard.us.udel.edu> Is it possible to set up dependencies between cluster services? That is, I have services A, B, C, and D. B, C, D can't run unless A is running, but B, C, and D are all independent of each other and I want to be able to control them individually, i.e. be able to start/stop (or rather, enable/disable) each without affecting the others. I know I could define them as dependent resources all in the same service, but then I can't have that independence between B, C, and D ... unless I'm missing something. Of course, the fall-back approach is that I just define each as a separate service and make the init scripts for B, C, and D check to see that A is running (probably via a clustat | grep "A's name" kludge), but I'm hoping there's a better way, since this is ugly as hell and will only work if the service consists entirely of init script resources. -g Greg Forte gforte at udel.edu IT - User Services University of Delaware 302-831-1982 Newark, DE From brilong at cisco.com Fri Feb 3 13:05:11 2006 From: brilong at cisco.com (Brian Long) Date: Fri, 03 Feb 2006 08:05:11 -0500 Subject: [Linux-cluster] Errata Mailing List? In-Reply-To: References: <1138898434.4441.14.camel@brilong-lnx> Message-ID: <1138971911.4472.12.camel@brilong-lnx> On Thu, 2006-02-02 at 22:03 +0100, Daniel EPEE LEA wrote: > Hi Brian, > I do not know about a list yet, but i know that if you have GFS > licenses or RHEL Entitlements you will receive daily errata notices > through your RHN account. That is how I get them. > > Hope this helps > > Daniel Daniel, I had forgotten about that since we've had it turned off for so long and prefer to stay subscribed to enterprise-watch-list. RHN only allows me to enable/disable email notifications; I wish I could select just certain channels. Oh well, I'll re-enable for now and work with RH on improvements. Thanks. /Brian/ > > On 2/2/06, Brian Long wrote: > > Hello, > > > > I just joined the list and was wondering if there is a mailman list > > similar to enterprise-watch-list that gives errata notices for GFS and > > RHCS. > > > > Thanks. > > > > /Brian/ > > -- > > Brian Long | | | > > IT Data Center Systems | .|||. .|||. > > Cisco Linux Developer | ..:|||||||:...:|||||||:.. > > Phone: (919) 392-7363 | C i s c o S y s t e m s > > > > -- > > Linux-cluster mailing list > > Linux-cluster at redhat.com > > https://www.redhat.com/mailman/listinfo/linux-cluster > > > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From lhh at redhat.com Fri Feb 3 15:18:28 2006 From: lhh at redhat.com (Lon Hohberger) Date: Fri, 03 Feb 2006 10:18:28 -0500 Subject: [Linux-cluster] /etc/cluster/cluster.conf In-Reply-To: <43E226F4.7030001@irisa.fr> References: <43E226F4.7030001@irisa.fr> Message-ID: <1138979908.5992.52.camel@ayanami.boston.redhat.com> On Thu, 2006-02-02 at 16:36 +0100, Anthony Assi wrote: > I am debugging a Performance problem we are having with our Cluster, > We have a 32 nodes cluster, using a GFS shared Volume, > I would like to know whether this is a good configuration of the > "cluster.conf" File: > > > > > fstype="gfs" mountpoint="/bdspecs" name="bdspecs" options="num_glockd=32"/> > fstype="gfs" mountpoint="/home/genouest" name="genouest" > options="num_glockd=32"/> > fstype="gfs" mountpoint="/home/symbiose" name="symbiose" > options="num_glockd=32"/> > fstype="gfs" mountpoint="/home/irisa" name="irisa" options="num_glockd=32"/> > fstype="gfs" mountpoint="/db" name="db" options="num_glockd=32"/> > fstype="gfs" mountpoint="/index" name="index" options="num_glockd=32"/> > > You really don't need rgmanager in this case. The above rgmanager configuration is just about equivalent to putting the GFS volumes in /etc/fstab -- Lon From lhh at redhat.com Fri Feb 3 15:39:47 2006 From: lhh at redhat.com (Lon Hohberger) Date: Fri, 03 Feb 2006 10:39:47 -0500 Subject: [Linux-cluster] dependencies between services In-Reply-To: <43E2948D.3060108@leopard.us.udel.edu> References: <43E2948D.3060108@leopard.us.udel.edu> Message-ID: <1138981187.5992.62.camel@ayanami.boston.redhat.com> On Thu, 2006-02-02 at 18:23 -0500, Greg Forte wrote: > Is it possible to set up dependencies between cluster services? That > is, I have services A, B, C, and D. B, C, D can't run unless A is > running, but B, C, and D are all independent of each other and I want to > be able to control them individually, i.e. be able to start/stop (or > rather, enable/disable) each without affecting the others. I know I > could define them as dependent resources all in the same service, but > then I can't have that independence between B, C, and D ... unless I'm > missing something. Not at the moment, but it should not be a difficult thing to add. Could you file a bugzilla about it? -- Lon From omer at faruk.net Sat Feb 4 20:21:22 2006 From: omer at faruk.net (Omer Faruk Sen) Date: Sat, 4 Feb 2006 22:21:22 +0200 (EET) Subject: [Linux-cluster] two nodes -> tree nodes In-Reply-To: <43E23012.5010405@redhat.com> References: <43E20FF9.4060406@linworx.cz> <43E23012.5010405@redhat.com> Message-ID: <62448.85.103.165.73.1139084482.squirrel@85.103.165.73> Then is it possible to create a cluster with 2 modes without setting two_node="1" ? Thus it will enable us not to reboot our servers when adding the third node. > Stanislav Pol??ek wrote: >> Hi everybody. I would like to add third node to the running cluster >> without restart of the cluster services. But when I try to join the >> cluster with the third node, the other two refuse the requests with the >> message "CMAN: join request from node requested, exceeds two node >> limit". My cluster.conf does not explicitly state the two_node >> parameter, but cman seems to run in that mode. What would be the right >> way to add the third node without restart of the cluster services on the >> other two? >> > > > two_node="1" Must have been in the cluster.conf file when the cluster was > formed - that's the only way the flag inside CMAN can get set. > > Once you have a 2 node cluster the only way to add a node is to remove all > nodes from the cluster and join them again without two_node set. > >>From 3 node upwards you can add nodes as you please - it's just the >> transition > from 2 to 3 that needs a restart because 2 nodes is a special case. > > -- > > patrick > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -- Omer Faruk Sen http://www.faruk.net From omer at faruk.net Sat Feb 4 20:25:04 2006 From: omer at faruk.net (Omer Faruk Sen) Date: Sat, 4 Feb 2006 22:25:04 +0200 (EET) Subject: [Linux-cluster] fencing and node relation? Message-ID: <52713.85.103.165.73.1139084704.squirrel@85.103.165.73> Hi, There is one issue that I don't grasp with fencing devices.. I will a 2 node cluster with dl140 (using fence_ipmilan) but what I don't understand is how will I relate a cluster node with the fencing device that this cluster node connected. For example I have node1 and node2 node1: 192.168.1.1 node2: 192.168.1.2 node1_ipmi: 192.168.1.253 node2_ipmi: 192.168.1.254 using system-config-cluster I can't see how to relate node1 and node1_ipmi so in the event of failure of node1 node2 can send node1_ipmi reboot message -- Omer Faruk Sen http://www.faruk.net From omer at faruk.net Sat Feb 4 21:00:57 2006 From: omer at faruk.net (Omer Faruk Sen) Date: Sat, 4 Feb 2006 23:00:57 +0200 (EET) Subject: [Linux-cluster] fencing and node relation? In-Reply-To: <52713.85.103.165.73.1139084704.squirrel@85.103.165.73> References: <52713.85.103.165.73.1139084704.squirrel@85.103.165.73> Message-ID: <59439.85.103.165.73.1139086857.squirrel@85.103.165.73> I have found my own answer: http://mirror.centos.org/centos/4/docs/html/rh-cs-en-4/s1-config-powercontroller.html#FIG-SOFT-PWRCTRLR Select the !!! member for which you want to configure a power controller connection !!! and click Manage Fencing For This Node. The Fence Configuration dialog box is displayed as shown in Figure 3-8. Also taking a look to usage.txt: it shows that fencing devices are IN the clusternode knob.. > Hi, > > There is one issue that I don't grasp with fencing devices.. I will a 2 > node cluster with dl140 (using fence_ipmilan) but what I don't understand > is how will I relate a cluster node with the fencing device that this > cluster node connected. > > > For example I have node1 and node2 > > node1: 192.168.1.1 > node2: 192.168.1.2 > node1_ipmi: 192.168.1.253 > node2_ipmi: 192.168.1.254 > > using system-config-cluster I can't see how to relate node1 and node1_ipmi > so in the event of failure of node1 node2 can send node1_ipmi reboot > message > > > > > -- > Omer Faruk Sen > http://www.faruk.net > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -- Omer Faruk Sen http://www.faruk.net From tristram at ubernet.co.nz Sun Feb 5 09:22:39 2006 From: tristram at ubernet.co.nz (Tristram Cheer) Date: Sun, 05 Feb 2006 22:22:39 +1300 Subject: [Linux-cluster] RGManager Hanging on shutdown Message-ID: <43E5C3DF.2000601@ubernet.co.nz> Hi folks, I'm having issues with the shutdown of my cluster, background on our cluster is this. We have 4 Compaq 8500r's with 8 x Xeon 800mhz with 3gb of Ram in each, we are using Xen 3.0.1 to run 16 VM's each with shared access via GFS to a GNBD Raid5 Export. The last remaining issue i have to work out before we look at Testing and moving to production is this, /etc/init.d/rgmanger hangs when shutting down any node in the cluster, all i get on console is "Waiting for cluster services to stop", i have to kill the process to shutdown the system, this happens on our VM or on Real hardware so i dont think the issue lies with Xen. Services running on each node are these root at edward:~# cat /proc/cluster/services Service Name GID LID State Code Fence Domain: "default" 1 2 run - [3 2 1 4 5 6 7 8 9] DLM Lock Space: "shared" 3 3 run - [3 2 9] GFS Mount Group: "shared" 4 4 run - [3 2 9] User: "usrm::manager" 2 5 run - [3 2 1 4 5 6 7 8 9] I cant figure out why this is happening, here is an strace from the shutdown Any help or tips to fixing this would be great root at edward:~# strace /etc/init.d/rgmanager stop execve("/etc/init.d/rgmanager", ["/etc/init.d/rgmanager", "stop"], [/* 17 vars */]) = 0 uname({sys="Linux", node="edward", ...}) = 0 brk(0) = 0x80eb000 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7f02000 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) old_mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7f00000 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=8432, ...}) = 0 old_mmap(NULL, 8432, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7efd000 close(3) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libncurses.so.5", O_RDONLY) = 3 read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\340\342"..., 512) = 512 fstat64(3, {st_mode=S_IFREG|0644, st_size=260524, ...}) = 0 old_mmap(NULL, 265868, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7ebc000 old_mmap(0xb7ef4000, 36864, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x37000) = 0xb7ef4000 close(3) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libdl.so.2", O_RDONLY) = 3 read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\240\v\0"..., 512) = 512 fstat64(3, {st_mode=S_IFREG|0644, st_size=8016, ...}) = 0 old_mmap(NULL, 10828, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7eb9000 old_mmap(0xb7ebb000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1000) = 0xb7ebb000 close(3) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libc.so.6", O_RDONLY) = 3 read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\306S\1"..., 512) = 512 fstat64(3, {st_mode=S_IFREG|0644, st_size=1131932, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7eb8000 old_mmap(NULL, 1141908, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7da1000 old_mmap(0xb7eb2000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x110000) = 0xb7eb2000 old_mmap(0xb7eb6000, 7316, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0xb7eb6000 close(3) = 0 munmap(0xb7efd000, 8432) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 open("/dev/tty", O_RDWR|O_NONBLOCK|O_LARGEFILE) = 3 close(3) = 0 open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=290448, ...}) = 0 mmap2(NULL, 290448, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7d5a000 close(3) = 0 brk(0) = 0x80eb000 brk(0x80ec000) = 0x80ec000 brk(0x80ed000) = 0x80ed000 brk(0x80ee000) = 0x80ee000 getuid32() = 0 getgid32() = 0 geteuid32() = 0 getegid32() = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 time(NULL) = 1139130793 brk(0x80ef000) = 0x80ef000 open("/etc/mtab", O_RDONLY) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=281, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7d59000 read(3, "/dev/hda1 / ext3 rw 0 0\nproc /pr"..., 4096) = 281 close(3) = 0 munmap(0xb7d59000, 4096) = 0 open("/proc/meminfo", O_RDONLY) = 3 fstat64(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7d59000 read(3, "MemTotal: 65728 kB\nMemFre"..., 1024) = 598 close(3) = 0 munmap(0xb7d59000, 4096) = 0 brk(0x80f0000) = 0x80f0000 rt_sigaction(SIGCHLD, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGCHLD, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGQUIT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGQUIT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigaction(SIGQUIT, {SIG_IGN}, {SIG_DFL}, 8) = 0 uname({sys="Linux", node="edward", ...}) = 0 brk(0x80f1000) = 0x80f1000 stat64("/root", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 stat64(".", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 getpid() = 23624 open("/usr/lib/gconv/gconv-modules.cache", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/lib/gconv/gconv-modules", O_RDONLY) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=45568, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7d59000 read(3, "# GNU libc iconv configuration.\n"..., 4096) = 4096 brk(0x80f2000) = 0x80f2000 read(3, "lias\tJS//\t\t\tJUS_I.B1.002//\nalias"..., 4096) = 4096 brk(0x80f3000) = 0x80f3000 brk(0x80f4000) = 0x80f4000 brk(0x80f5000) = 0x80f5000 read(3, "ule\tINTERNAL\t\tISO-8859-3//\t\tISO8"..., 4096) = 4096 brk(0x80f6000) = 0x80f6000 brk(0x80f7000) = 0x80f7000 brk(0x80f8000) = 0x80f8000 read(3, "lias\tISO-IR-199//\t\tISO-8859-14//"..., 4096) = 4096 brk(0x80f9000) = 0x80f9000 brk(0x80fa000) = 0x80fa000 read(3, "\t\tto\t\t\tmodule\t\tcost\nalias\tCSEBCD"..., 4096) = 4096 brk(0x80fb000) = 0x80fb000 brk(0x80fc000) = 0x80fc000 brk(0x80fd000) = 0x80fd000 read(3, "ule\t\tcost\nalias\tCP284//\t\t\tIBM284"..., 4096) = 4096 brk(0x80fe000) = 0x80fe000 brk(0x80ff000) = 0x80ff000 brk(0x8100000) = 0x8100000 read(3, "lias\tCP864//\t\t\tIBM864//\nalias\t86"..., 4096) = 4096 brk(0x8101000) = 0x8101000 brk(0x8102000) = 0x8102000 brk(0x8103000) = 0x8103000 read(3, "module\tIBM937//\t\tINTERNAL\t\tIBM93"..., 4096) = 4096 brk(0x8104000) = 0x8104000 brk(0x8105000) = 0x8105000 brk(0x8106000) = 0x8106000 read(3, "\tEUC-JP//\nalias\tUJIS//\t\t\tEUC-JP/"..., 4096) = 4096 brk(0x8107000) = 0x8107000 brk(0x8108000) = 0x8108000 read(3, "module\t\tcost\nalias\tISO-IR-143//\t"..., 4096) = 4096 brk(0x8109000) = 0x8109000 brk(0x810a000) = 0x810a000 brk(0x810b000) = 0x810b000 read(3, "-BOX//\nmodule\tISO_10367-BOX//\t\tI"..., 4096) = 4096 brk(0x810c000) = 0x810c000 brk(0x810d000) = 0x810d000 brk(0x810e000) = 0x810e000 read(3, "module\tINTERNAL\t\tEUC-JISX0213//\t"..., 4096) = 512 read(3, "", 4096) = 0 close(3) = 0 munmap(0xb7d59000, 4096) = 0 brk(0x810f000) = 0x810f000 open("/usr/lib/gconv/ISO8859-1.so", O_RDONLY) = 3 read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0`\4\0\000"..., 512) = 512 fstat64(3, {st_mode=S_IFREG|0644, st_size=5576, ...}) = 0 brk(0x8110000) = 0x8110000 old_mmap(NULL, 4368, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7d58000 old_mmap(0xb7d59000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1000) = 0xb7d59000 close(3) = 0 getppid() = 23623 getpgrp() = 23623 rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 open("/etc/init.d/rgmanager", O_RDONLY|O_LARGEFILE) = 3 ioctl(3, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbff1698c) = -1 ENOTTY (Inappropriate ioctl for device) _llseek(3, 0, [0], SEEK_CUR) = 0 read(3, "#!/bin/sh\n\nPATH=/usr/local/sbin:"..., 80) = 80 _llseek(3, 0, [0], SEEK_SET) = 0 getrlimit(RLIMIT_NOFILE, {rlim_cur=1024, rlim_max=1024}) = 0 dup2(3, 255) = 255 close(3) = 0 fcntl64(255, F_SETFD, FD_CLOEXEC) = 0 fcntl64(255, F_GETFL) = 0x8000 (flags O_RDONLY|O_LARGEFILE) fstat64(255, {st_mode=S_IFREG|0755, st_size=1319, ...}) = 0 _llseek(255, 0, [0], SEEK_CUR) = 0 brk(0x8111000) = 0x8111000 brk(0x8112000) = 0x8112000 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 read(255, "#!/bin/sh\n\nPATH=/usr/local/sbin:"..., 1319) = 1319 brk(0x8113000) = 0x8113000 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 stat64("/etc/default/rgmanager", {st_mode=S_IFREG|0644, st_size=19, ...}) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/etc/default/rgmanager", {st_mode=S_IFREG|0644, st_size=19, ...}) = 0 open("/etc/default/rgmanager", O_RDONLY|O_LARGEFILE) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=19, ...}) = 0 read(3, "# RGMGR_OPTIONS=\"\"\n", 19) = 19 close(3) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 brk(0x8114000) = 0x8114000 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 brk(0x8115000) = 0x8115000 brk(0x8116000) = 0x8116000 brk(0x8117000) = 0x8117000 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64(".", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 stat64("/usr/local/sbin/clulog", 0xbff14088) = -1 ENOENT (No such file or directory) stat64("/usr/local/bin/clulog", 0xbff14088) = -1 ENOENT (No such file or directory) stat64("/sbin/clulog", 0xbff14088) = -1 ENOENT (No such file or directory) stat64("/bin/clulog", 0xbff14088) = -1 ENOENT (No such file or directory) stat64("/usr/sbin/clulog", {st_mode=S_IFREG|0755, st_size=6632, ...}) = 0 stat64("/usr/sbin/clulog", {st_mode=S_IFREG|0755, st_size=6632, ...}) = 0 brk(0x8118000) = 0x8118000 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23625 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23625 waitpid(-1, 0xbff13dbc, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 write(1, "Stopping cluster service manager"..., 34Stopping cluster service manager: ) = 34 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64(".", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 stat64("/usr/local/sbin/start-stop-daemon", 0xbff15128) = -1 ENOENT (No such file or directory) stat64("/usr/local/bin/start-stop-daemon", 0xbff15128) = -1 ENOENT (No such file or directory) stat64("/sbin/start-stop-daemon", {st_mode=S_IFREG|0755, st_size=18520, ...}) = 0 stat64("/sbin/start-stop-daemon", {st_mode=S_IFREG|0755, st_size=18520, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23626 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23626 waitpid(-1, 0xbff14e5c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23627 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23627 waitpid(-1, 0xbff1412c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 write(1, "Waiting for services to stop: ", 30Waiting for services to stop: ) = 30 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23628 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23628 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64(".", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 stat64("/usr/local/sbin/sleep", 0xbff13ce8) = -1 ENOENT (No such file or directory) stat64("/usr/local/bin/sleep", 0xbff13ce8) = -1 ENOENT (No such file or directory) stat64("/sbin/sleep", 0xbff13ce8) = -1 ENOENT (No such file or directory) stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23629 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23629 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23630 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23630 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23631 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23631 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23632 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23632 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23633 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23633 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23637 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23637 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23638 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23638 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23639 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23639 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23640 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23640 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23641 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23641 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23642 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23642 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23643 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23643 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23644 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 23644 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, 0xbff13a9c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 pipe([3, 4]) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23645 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- waitpid(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], WNOHANG) = 23645 waitpid(-1, 0xbff1374c, WNOHANG) = -1 ECHILD (No child processes) sigreturn() = ? (mask now []) rt_sigaction(SIGCHLD, {0x8078e75, [], 0}, {0x8078e75, [], 0}, 8) = 0 close(4) = 0 read(3, "4519 4478\n", 128) = 10 read(3, "", 128) = 0 close(3) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {0x8077b64, [], 0}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 stat64("/bin/sleep", {st_mode=S_IFREG|0755, st_size=13920, ...}) = 0 rt_sigprocmask(SIG_BLOCK, [INT CHLD], [], 8) = 0 fork() = 23646 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0 rt_sigaction(SIGINT, {0x8077b64, [], 0}, {SIG_DFL}, 8) = 0 waitpid(-1, Tristram Cheer From omer at faruk.net Sun Feb 5 17:41:27 2006 From: omer at faruk.net (Omer Faruk Sen) Date: Sun, 5 Feb 2006 19:41:27 +0200 (EET) Subject: [Linux-cluster] #12: RG xxcluster failed to stop; intervention required Message-ID: <64526.85.103.165.73.1139161287.squirrel@85.103.165.73> Hi, I have done a simple test but redhat-cluster can't survive it. I have a 2 node cluster I have edited httpd.conf so it can't start (adding a few characters to httpd.conf) and on node2 I have made a /usr/sbin/clusvcadm -r cluster -m clu2 it move resources to clu2 perfectly but when I corrected httpd.conf on clu1 and issue /usr/sbin/clusvcadm -r ggcluster -m clu1 on node2 I get following errors: Feb 5 16:25:19 clu1 clurgmgrd[2256]: stop on script "apache" returned 1 (generic error) Feb 5 16:25:19 clu1 clurgmgrd[2256]: #12: RG xxcluster failed to stop; intervention required Feb 5 16:25:19 clu1 clurgmgrd[2256]: Service xxcluster is failed It shows that httpd service can't get started but when I manually start apache it starts without problem .. Can someone tell me what I am missing. It is a simple test and redhat cluster didn't pass it. I use 2 dl140 using fence_ipmilan. I can manually stop start and reboot server using fence_ipmilan command. And here is my cluster.conf: