From jamescyriac76 at gmail.com Sun Dec 2 03:34:52 2012 From: jamescyriac76 at gmail.com (james cyriac) Date: Sun, 2 Dec 2012 07:34:52 +0400 Subject: [Linux-cluster] Normal startup vs startup due to failover on cluster node - can they be distinguished? In-Reply-To: References: Message-ID: Hi all i have 2 node cluster redaht 6.2.but after rebooting the node both not joining to cluster.i have to join manually.i have to add on start up any script? Thanks james On Fri, Nov 23, 2012 at 9:25 AM, Parvez Shaikh wrote: > Hi experts, > > I am using Red Hat Cluster available on RHEL 5.5. And it doesn't have any > inbuilt mechanism to generate SNMP traps in failures of resources or > failover of services from one node to another. > > I have a script agent, which starts, stops and checks status of my > application. Is it possible that in a script resource - to distinguish > between normal startup of service / resource vs startup of service/resource > in response to failover / failure handling? Doing so would help me write > code to generate alarms if startup of service / resource (in my case a > process) is due to failover (not normal startup). > > Further is it possible to get information such as cause of failure(leading > to failover), and previous cluster node on which service / resource was > running(prior to failover)? > > This would help to provide as much information as possible in traps > > Thanks, > Parvez > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dlcarvalho at gmail.com Mon Dec 3 18:29:11 2012 From: dlcarvalho at gmail.com (Daniel Lopes de Carvalho) Date: Mon, 3 Dec 2012 16:29:11 -0200 Subject: [Linux-cluster] Pacemaker+DRBD+GFS2 multiple filesystems. Message-ID: Hi. I?m new to pacemaker and I?m looking for a way to configure pacemaker with drbd and gfs2 with two filesystems (/home and /export). My issue is under paceaker configuration resources. I can?t get two filesystem resource working on the same time. I always get an error on the second fs. Could someone guide me with this configurations? Thanks Best Regards. -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew at beekhof.net Wed Dec 5 00:39:40 2012 From: andrew at beekhof.net (Andrew Beekhof) Date: Wed, 5 Dec 2012 11:39:40 +1100 Subject: [Linux-cluster] Pacemaker+DRBD+GFS2 multiple filesystems. In-Reply-To: References: Message-ID: What does your config look like? What was the error you get? Logs? On Tuesday, December 4, 2012, Daniel Lopes de Carvalho wrote: > Hi. > > I?m new to pacemaker and I?m looking for a way to configure pacemaker with > drbd and gfs2 with two filesystems (/home and /export). > > My issue is under paceaker configuration resources. I can?t get two > filesystem resource working on the same time. I always get an error on the > second fs. > > Could someone guide me with this configurations? > > Thanks > > Best Regards. > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashish.ext123 at gmail.com Wed Dec 5 13:52:15 2012 From: ashish.ext123 at gmail.com (Ashish G) Date: Wed, 5 Dec 2012 19:22:15 +0530 Subject: [Linux-cluster] CCSD ipv6 bind failures Message-ID: hi Experts, I has few question on ccsd: 1. what is the purpose of ccsd listening on ipv4 and ipv6 addresses as follows in my 2 node HA setup? We do not use IPv6 in our setup. netstat -antp |grep ccsd tcp 0 0 0.0.0.0:50008 0.0.0.0:* LISTEN 18170/ccsd *tcp 0 0 ::1:50006 :::* LISTEN 18170/ccsd* ** *CCSD is invoked without any arguments: * *root 18170 1 0 14:12 ? 00:00:00 /sbin/ccsd* 2. Is it not sufficient to use IPv4 only by CCSD? 3. Sometimes on node reboot we see messages as follows: "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL pointer: Address already in use" Does anyone know when is this problem seen? 4. If I invoke CCSD with "-4" option, will I get rid of the above message? since ipv6 code path will not hit. *root 461 1 0 14:28 ? 00:00:00 /sbin/ccsd -4* netstat -antp |grep ccsd *tcp 0 0 127.0.0.1:50006 0.0.0.0:* LISTEN 461/ccsd* tcp 0 0 0.0.0.0:50008 0.0.0.0:* LISTEN 461/ccsd Regards, Ashish -------------- next part -------------- An HTML attachment was scrubbed... URL: From felipe.o.gutierrez at gmail.com Fri Dec 7 19:04:49 2012 From: felipe.o.gutierrez at gmail.com (Felipe Gutierrez) Date: Fri, 7 Dec 2012 16:04:49 -0300 Subject: [Linux-cluster] Reconecting node using heartbeat Message-ID: Hi everyone, I am using heartbeat on Ubuntu. I have two nodes configured and both have two network board. So I can connect them with a crossover cable. When I disconnect one node, the heartbeat detects it. But when I reconnect it, the heartbeat doesn't recognize. Does anynoy know why? Thanks in advence. # cat ha.cf logfacility local0 udpport 694 keepalive 1 deadtime 10 warntime 3 initdead 20 bcast eth1 auto_failback on watchdog /dev/watchdog debugfile /var/log/ha-debug node cloud4 node cloud11 crm on # crm_mon -1 -V crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Resource start-up disabled since no STONITH resources have been defined crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Either configure some or disable STONITH with the stonith-enabled option crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: NOTE: Clusters with shared data need STONITH to ensure data integrity ============ Last updated: Fri Dec 7 18:04:58 2012 Last change: Fri Dec 7 17:42:08 2012 via crmd on cloud4 Stack: Heartbeat Current DC: cloud4 (01f645d4-c8d5-4b2a-8c44-5727101d56c5) - partition with quorum Version: 1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c 2 Nodes configured, unknown expected votes 0 Resources configured. ============ Node cloud11 (193a0634-90b5-4c1b-aed1-b987dbae188e): UNCLEAN (offline) Online: [ cloud4 ] -- *-- -- Felipe Oliveira Gutierrez -- Felipe.o.Gutierrez at gmail.com -- https://sites.google.com/site/lipe82/Home/diaadia* -------------- next part -------------- An HTML attachment was scrubbed... URL: From lists at alteeve.ca Fri Dec 7 19:12:34 2012 From: lists at alteeve.ca (Digimer) Date: Fri, 07 Dec 2012 14:12:34 -0500 Subject: [Linux-cluster] Reconecting node using heartbeat In-Reply-To: References: Message-ID: <50C23FA2.2040409@alteeve.ca> Any reason for not using corosync? Heartbeat hasn't been developed in some time and there are no plans to restart development in the future. On 12/07/2012 02:04 PM, Felipe Gutierrez wrote: > Hi everyone, > > > I am using heartbeat on Ubuntu. I have two nodes configured and both > have two network board. So I can connect them with a crossover cable. > When I disconnect one node, the heartbeat detects it. But when I > reconnect it, the heartbeat doesn't recognize. Does anynoy know why? > > Thanks in advence. > > # cat ha.cf > logfacility local0 > udpport 694 > keepalive 1 > deadtime 10 > warntime 3 > initdead 20 > bcast eth1 > auto_failback on > watchdog /dev/watchdog > debugfile /var/log/ha-debug > node cloud4 > node cloud11 > crm on > > # crm_mon -1 -V > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Resource > start-up disabled since no STONITH resources have been defined > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Either > configure some or disable STONITH with the stonith-enabled option > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: NOTE: > Clusters with shared data need STONITH to ensure data integrity > ============ > Last updated: Fri Dec 7 18:04:58 2012 > Last change: Fri Dec 7 17:42:08 2012 via crmd on cloud4 > Stack: Heartbeat > Current DC: cloud4 (01f645d4-c8d5-4b2a-8c44-5727101d56c5) - partition > with quorum > Version: 1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c > 2 Nodes configured, unknown expected votes > 0 Resources configured. > ============ > > Node cloud11 (193a0634-90b5-4c1b-aed1-b987dbae188e): UNCLEAN (offline) > Online: [ cloud4 ] > > -- > *-- > -- Felipe Oliveira Gutierrez > -- Felipe.o.Gutierrez at gmail.com > -- https://sites.google.com/site/lipe82/Home/diaadia* > > -- Digimer Papers and Projects: https://alteeve.ca/w/ What if the cure for cancer is trapped in the mind of a person without access to education? From lists at alteeve.ca Fri Dec 7 19:28:57 2012 From: lists at alteeve.ca (Digimer) Date: Fri, 07 Dec 2012 14:28:57 -0500 Subject: [Linux-cluster] Reconecting node using heartbeat In-Reply-To: References: <50C23FA2.2040409@alteeve.ca> Message-ID: <50C24379.9000306@alteeve.ca> On 12/07/2012 02:25 PM, Felipe Gutierrez wrote: > Hi Digimer, > > Thanks for reply. > I just found this site > > to help to cofigure Xen+DRBD+Corosync+OCFS2. > For Heartbeat I found more. > > But after you say that I will restart my sistem using corosync. Do tou > have others sites to help? > > Thanks. > Felipe I don't know what your goal is, so I can't say really. In general though; If you want Red Hat support; Use corosync + cman + rgmanager - https://alteeve.ca/w/2-Node_Red_Hat_KVM_Cluster_Tutorial Otherwise, use corosync + pacemaker (Red Hat will switch to pacemaker in RHEL 7, pacemaker is the future). Look for "Clusters From Scratch"; - http://clusterlabs.org/doc/ -- Digimer Papers and Projects: https://alteeve.ca/w/ What if the cure for cancer is trapped in the mind of a person without access to education? From felipe.o.gutierrez at gmail.com Fri Dec 7 19:25:05 2012 From: felipe.o.gutierrez at gmail.com (Felipe Gutierrez) Date: Fri, 7 Dec 2012 16:25:05 -0300 Subject: [Linux-cluster] Reconecting node using heartbeat In-Reply-To: <50C23FA2.2040409@alteeve.ca> References: <50C23FA2.2040409@alteeve.ca> Message-ID: Hi Digimer, Thanks for reply. I just found this siteto help to cofigure Xen+DRBD+Corosync+OCFS2. For Heartbeat I found more. But after you say that I will restart my sistem using corosync. Do tou have others sites to help? Thanks. Felipe On Fri, Dec 7, 2012 at 5:12 PM, Digimer wrote: > Any reason for not using corosync? Heartbeat hasn't been developed in > some time and there are no plans to restart development in the future. > > On 12/07/2012 02:04 PM, Felipe Gutierrez wrote: > > Hi everyone, > > > > > > I am using heartbeat on Ubuntu. I have two nodes configured and both > > have two network board. So I can connect them with a crossover cable. > > When I disconnect one node, the heartbeat detects it. But when I > > reconnect it, the heartbeat doesn't recognize. Does anynoy know why? > > > > Thanks in advence. > > > > # cat ha.cf > > logfacility local0 > > udpport 694 > > keepalive 1 > > deadtime 10 > > warntime 3 > > initdead 20 > > bcast eth1 > > auto_failback on > > watchdog /dev/watchdog > > debugfile /var/log/ha-debug > > node cloud4 > > node cloud11 > > crm on > > > > # crm_mon -1 -V > > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Resource > > start-up disabled since no STONITH resources have been defined > > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: Either > > configure some or disable STONITH with the stonith-enabled option > > crm_mon[5370]: 2012/12/07_18:04:58 ERROR: unpack_resources: NOTE: > > Clusters with shared data need STONITH to ensure data integrity > > ============ > > Last updated: Fri Dec 7 18:04:58 2012 > > Last change: Fri Dec 7 17:42:08 2012 via crmd on cloud4 > > Stack: Heartbeat > > Current DC: cloud4 (01f645d4-c8d5-4b2a-8c44-5727101d56c5) - partition > > with quorum > > Version: 1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c > > 2 Nodes configured, unknown expected votes > > 0 Resources configured. > > ============ > > > > Node cloud11 (193a0634-90b5-4c1b-aed1-b987dbae188e): UNCLEAN (offline) > > Online: [ cloud4 ] > > > > -- > > *-- > > -- Felipe Oliveira Gutierrez > > -- Felipe.o.Gutierrez at gmail.com > > -- https://sites.google.com/site/lipe82/Home/diaadia* > > > > > > > -- > Digimer > Papers and Projects: https://alteeve.ca/w/ > What if the cure for cancer is trapped in the mind of a person without > access to education? > -- *-- -- Felipe Oliveira Gutierrez -- Felipe.o.Gutierrez at gmail.com -- https://sites.google.com/site/lipe82/Home/diaadia* -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Mon Dec 10 08:02:59 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 10 Dec 2012 09:02:59 +0100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes Message-ID: Hello, I was evaluating migration from a rhel 5.8 two node cluster with services based on ha-lvm to a clvmd one. I'm using netapp storage with snapshot technology. I quiesce db, then make a snapshot and then a third server connects to the flashcopy volume, with "snapdrive connect" command and makes a backup of the db. Passing to clvmd, the command fails because it tries to do a vgimport and returns error because it is a clustered volume. The operation is atomic, so I didn't find a way to only present the lun and run manual command... Any suggestion about working at origin and temporarily disable cluster flag before snapshot? Or other ones with similar problems? Thanks in advance Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From lwj at gildata.com Mon Dec 10 09:19:06 2012 From: lwj at gildata.com (lwj) Date: Mon, 10 Dec 2012 17:19:06 +0800 Subject: [Linux-cluster] (no subject) References: Message-ID: <201212101719063754087@gildata.com> From ???(Liu Wei Jie) ?????????????? (86)2160897890 From: Gianluca Cecchi Date: 2012-12-10 16:02 To: linux clustering Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes Hello, I was evaluating migration from a rhel 5.8 two node cluster with services based on ha-lvm to a clvmd one. I'm using netapp storage with snapshot technology. I quiesce db, then make a snapshot and then a third server connects to the flashcopy volume, with "snapdrive connect" command and makes a backup of the db. Passing to clvmd, the command fails because it tries to do a vgimport and returns error because it is a clustered volume. The operation is atomic, so I didn't find a way to only present the lun and run manual command... Any suggestion about working at origin and temporarily disable cluster flag before snapshot? Or other ones with similar problems? Thanks in advance Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From songyu555 at gmail.com Mon Dec 10 10:02:19 2012 From: songyu555 at gmail.com (Yu) Date: Mon, 10 Dec 2012 21:02:19 +1100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes In-Reply-To: References: Message-ID: <0F85FC46-9554-42BF-958F-06F47AC05293@gmail.com> Vgchange -cn vgname Can disable cluster flag Regards Yu On 10/12/2012, at 19:02, Gianluca Cecchi wrote: > Hello, I was evaluating migration from a rhel 5.8 two node cluster with services based on ha-lvm to a clvmd one. > I'm using netapp storage with snapshot technology. > I quiesce db, then make a snapshot and then a third server connects to the flashcopy volume, with "snapdrive connect" command and makes a backup of the db. > Passing to clvmd, the command fails because it tries to do a vgimport and returns error because it is a clustered volume. > The operation is atomic, so I didn't find a way to only present the lun and run manual command... > Any suggestion about working at origin and temporarily disable cluster flag before snapshot? > Or other ones with similar problems? > Thanks in advance > Gianluca > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Mon Dec 10 12:13:59 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 10 Dec 2012 13:13:59 +0100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes In-Reply-To: References: Message-ID: On Mon, 10 Dec 2012 21:02:19 +1100 Yu wrote: > vgchange -cn vgname > Can disable cluster flag > Regards > Yu Ok, thanks, I thought about this option too and it seems it works. More testing in place. In this cluster I have many services and each one is bundled with its own VGs, so that each node is active for that particular service. To use CLVMD is comfortable when I need to add LUNs or resize pre-existing ones. I had a doubt if running "vgchange -cn" when both nodes are online could cause any problem so I was not sure... What does it happen with node 2 when I run this command on node 1? "SImply" that node cannot online the VG any more or worse it can online it without safety and cause data corruption? What the best behaviour to avoid data corruption for the few seconds I need to make a snapshot of a not-clustered VG? Gianluca From songyu555 at gmail.com Mon Dec 10 12:36:51 2012 From: songyu555 at gmail.com (yu song) Date: Mon, 10 Dec 2012 23:36:51 +1100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes In-Reply-To: References: Message-ID: ok..firstly confirm whether you are using HA-LVM, not GFS/GFS2 .. correct? if ha-lvm, your vg should only be imported/mounted on one node, instead of multiple nodes. if you run this command on node1, assuming vg1, you still can use "vgchange -cy vg1" on node2 to import the vg and use it on node2. as long as you don't mount volumes of same vg ( still assuming you are not using gfs/gfs2) on multiple nodes at the same time.. you won't corrupt data. snapshot or backup is always an essential part when you do migration. Yu On Mon, Dec 10, 2012 at 11:13 PM, Gianluca Cecchi wrote: > On Mon, 10 Dec 2012 21:02:19 +1100 Yu wrote: > > vgchange -cn vgname > > Can disable cluster flag > > Regards > > Yu > > Ok, thanks, I thought about this option too and it seems it works. > More testing in place. > In this cluster I have many services and each one is bundled with its > own VGs, so that each node is active for that particular service. > To use CLVMD is comfortable when I need to add LUNs or resize pre-existing > ones. > I had a doubt if running "vgchange -cn" when both nodes are online > could cause any problem so I was not sure... > What does it happen with node 2 when I run this command on node 1? > "SImply" that node cannot online the VG any more or worse it can > online it without safety and cause data corruption? > What the best behaviour to avoid data corruption for the few seconds I > need to make a snapshot of a not-clustered VG? > > Gianluca > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Mon Dec 10 14:54:19 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 10 Dec 2012 15:54:19 +0100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes In-Reply-To: References: Message-ID: On Mon, 10 Dec 2012 23:36:51 +1100 yu song wrote: > ok..firstly confirm whether you are using HA-LVM, not GFS/GFS2 .. correct? > if ha-lvm, your vg should only be imported/mounted on one node, instead of multiple nodes. In my non-CLVMD configuration I use HA-LVM with these entries in cluster.conf ... .... and tagging in lvm.conf (volume_list directive) With CLVMD and the config I'm testig with snapshots and "vgchange -cn" I still don't use any GFSx file systems, only plain ext3, mounted only on the node carrying on the service, and my configuration simply doesn't contain the lvm part: ... .... Using locking_type=3 in lvm.conf while not using any tagging (volume_list directive) Any comments about that? From songyu555 at gmail.com Tue Dec 11 02:51:29 2012 From: songyu555 at gmail.com (yu song) Date: Tue, 11 Dec 2012 13:51:29 +1100 Subject: [Linux-cluster] Netapp snapdrive connect doesn't like clustered volumes In-Reply-To: References: Message-ID: lvm part still needs to be specified in the service section if you want to follow a proper way. locking_type =3 is correct.. you can use lvmconf --enable-cluster to enable clvmd, which actually does change locking type to 3 anyway. when you create a new vg controlled by clvmd.. after pvcreate/vgcreate/lvcreate/mkfs.ext3... don't forget to run lvchange -an vg0/ha_lv nothing special about tagging is requires as far as I am aware.. you might need to look at filter section to ensure the volumes are not blacklisted. On Tue, Dec 11, 2012 at 1:54 AM, Gianluca Cecchi wrote: > On Mon, 10 Dec 2012 23:36:51 +1100 yu song wrote: > > > ok..firstly confirm whether you are using HA-LVM, not GFS/GFS2 .. > correct? > > if ha-lvm, your vg should only be imported/mounted on one node, instead > of multiple nodes. > > In my non-CLVMD configuration I use HA-LVM with these entries in > cluster.conf > > > > fsid="50001" fstype="ext3" mountpoint="/my_mnt" name="MYFS" options="" > self_fence="1"/> > > ... > > > > > > > .... > > and tagging in lvm.conf (volume_list directive) > > With CLVMD and the config I'm testig with snapshots and "vgchange -cn" > I still don't use any GFSx file systems, only plain ext3, mounted only > on the node carrying on the service, and my configuration simply > doesn't contain the lvm part: > > > fsid="50001" fstype="ext3" mountpoint="/my_mnt" name="MYFS" options="" > self_fence="1"/> > > ... > > > > > > .... > Using locking_type=3 in lvm.conf while not using any tagging > (volume_list directive) > > Any comments about that? > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -------------- next part -------------- An HTML attachment was scrubbed... URL: From rossnick-lists at cybercat.ca Wed Dec 12 18:10:39 2012 From: rossnick-lists at cybercat.ca (Nicolas Ross) Date: Wed, 12 Dec 2012 13:10:39 -0500 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. Message-ID: <50C8C89F.9080200@cybercat.ca> Short question : How do I do it ? Longer explainaition : We got a VG with 5 PV located on a Fiber channel raid enclosure that is used to host GFS partiions among other things. I would like to move the extent for on LV from on PV to another within the same VG. When trying to do pvmove -n LVName oldPV newPV, I get : Cannot move in clustered VG VGa, clustered mirror (cmirror) not detected and LVs are activated non-exclusively. (VGa is my VG). So, I suppose I need to activate exclusively the LV to do so ? Is it safe to do it with clvmd ? Regards, From hal at elizium.za.net Wed Dec 12 18:39:37 2012 From: hal at elizium.za.net (Hugo Lombard) Date: Wed, 12 Dec 2012 20:39:37 +0200 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <50C8C89F.9080200@cybercat.ca> References: <50C8C89F.9080200@cybercat.ca> Message-ID: <20121212183937.GI14097@squishy.elizium.za.net> On Wed, Dec 12, 2012 at 01:10:39PM -0500, Nicolas Ross wrote: > Short question : How do I do it ? > > Longer explainaition : We got a VG with 5 PV located on a Fiber channel > raid enclosure that is used to host GFS partiions among other things. I > would like to move the extent for on LV from on PV to another within the > same VG. > > When trying to do pvmove -n LVName oldPV newPV, I get : > > Cannot move in clustered VG VGa, clustered mirror (cmirror) not detected > and LVs are activated non-exclusively. > > (VGa is my VG). > > So, I suppose I need to activate exclusively the LV to do so ? Is it > safe to do it with clvmd ? > AFAICR on CentOS 5 I had to install cmirror and the applicable kmod-cmirror, make sure the cmirror service was started, and then pvmove did it's thing. -- Hugo Lombard From rossnick-lists at cybercat.ca Wed Dec 12 19:30:57 2012 From: rossnick-lists at cybercat.ca (Nicolas Ross) Date: Wed, 12 Dec 2012 14:30:57 -0500 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <20121212183937.GI14097@squishy.elizium.za.net> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> Message-ID: <50C8DB71.7010009@cybercat.ca> > AFAICR on CentOS 5 I had to install cmirror and the applicable > kmod-cmirror, make sure the cmirror service was started, and then pvmove > did it's thing. Do I need to start it on all nodes ? For now, it's not installed at all. From hal at elizium.za.net Wed Dec 12 20:29:05 2012 From: hal at elizium.za.net (Hugo Lombard) Date: Wed, 12 Dec 2012 22:29:05 +0200 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <50C8DB71.7010009@cybercat.ca> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> Message-ID: <20121212202905.GJ14097@squishy.elizium.za.net> On Wed, Dec 12, 2012 at 02:30:57PM -0500, Nicolas Ross wrote: > > AFAICR on CentOS 5 I had to install cmirror and the applicable > > kmod-cmirror, make sure the cmirror service was started, and then pvmove > > did it's thing. > > Do I need to start it on all nodes ? For now, it's not installed at all. > Yes, I think you have to. -- Hugo Lombard From rossnick-lists at cybercat.ca Thu Dec 13 02:46:15 2012 From: rossnick-lists at cybercat.ca (Nicolas Ross) Date: Wed, 12 Dec 2012 21:46:15 -0500 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <20121212202905.GJ14097@squishy.elizium.za.net> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> Message-ID: <50C94177.8070706@cybercat.ca> Hugo Lombard a ?crit : > On Wed, Dec 12, 2012 at 02:30:57PM -0500, Nicolas Ross wrote: >>> AFAICR on CentOS 5 I had to install cmirror and the applicable >>> kmod-cmirror, make sure the cmirror service was started, and then pvmove >>> did it's thing. >> Do I need to start it on all nodes ? For now, it's not installed at all. >> > Yes, I think you have to. > I can't still do it... I had to deactivate the LV, and activate it exclusivly on one node, and issue the pvmove, and I still get : Error locking on node node206.lan: device-mapper: create ioctl on VGa-pvmove0 failed: Device or resource busy Failed to suspend logicalvolumename ABORTING: Volume group metadata update failed. (first_time: 1) Regards, From songyu555 at gmail.com Thu Dec 13 03:26:26 2012 From: songyu555 at gmail.com (Yu) Date: Thu, 13 Dec 2012 14:26:26 +1100 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <50C94177.8070706@cybercat.ca> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> <50C94177.8070706@cybercat.ca> Message-ID: <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> Alternatively, you might try to import vg with lickproto=lock_nolock, then do pvmove. Yu On 13/12/2012, at 13:46, Nicolas Ross wrote: > Hugo Lombard a ?crit : >> On Wed, Dec 12, 2012 at 02:30:57PM -0500, Nicolas Ross wrote: >>>> AFAICR on CentOS 5 I had to install cmirror and the applicable >>>> kmod-cmirror, make sure the cmirror service was started, and then pvmove >>>> did it's thing. >>> Do I need to start it on all nodes ? For now, it's not installed at all. >> Yes, I think you have to. > I can't still do it... > > I had to deactivate the LV, and activate it exclusivly on one node, and issue the pvmove, and I still get : > > Error locking on node node206.lan: device-mapper: create ioctl on VGa-pvmove0 failed: Device or resource busy > Failed to suspend logicalvolumename > ABORTING: Volume group metadata update failed. (first_time: 1) > > Regards, > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster From corey.kovacs at gmail.com Thu Dec 13 03:58:12 2012 From: corey.kovacs at gmail.com (Corey Kovacs) Date: Wed, 12 Dec 2012 20:58:12 -0700 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> <50C94177.8070706@cybercat.ca> <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> Message-ID: Cant you just do a "vgchange -aey" to put the vg into exclusive mode? Thats what used to be the way to do it. Its been a while though. On Dec 12, 2012 8:30 PM, "Yu" wrote: > Alternatively, you might try to import vg with lickproto=lock_nolock, then > do pvmove. > > Yu > > On 13/12/2012, at 13:46, Nicolas Ross wrote: > > > Hugo Lombard a ?crit : > >> On Wed, Dec 12, 2012 at 02:30:57PM -0500, Nicolas Ross wrote: > >>>> AFAICR on CentOS 5 I had to install cmirror and the applicable > >>>> kmod-cmirror, make sure the cmirror service was started, and then > pvmove > >>>> did it's thing. > >>> Do I need to start it on all nodes ? For now, it's not installed at > all. > >> Yes, I think you have to. > > I can't still do it... > > > > I had to deactivate the LV, and activate it exclusivly on one node, and > issue the pvmove, and I still get : > > > > Error locking on node node206.lan: device-mapper: create ioctl on > VGa-pvmove0 failed: Device or resource busy > > Failed to suspend logicalvolumename > > ABORTING: Volume group metadata update failed. (first_time: 1) > > > > Regards, > > > > -- > > 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 -------------- next part -------------- An HTML attachment was scrubbed... URL: From rossnick-lists at cybercat.ca Thu Dec 13 14:33:59 2012 From: rossnick-lists at cybercat.ca (Nicolas Ross) Date: Thu, 13 Dec 2012 09:33:59 -0500 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> <50C94177.8070706@cybercat.ca> <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> Message-ID: <50C9E757.4080005@cybercat.ca> > Cant you just do a "vgchange -aey" to put the vg into exclusive mode? > Thats what used to be the way to do it. Its been a while though. > > On Dec 12, 2012 8:30 PM, "Yu" > wrote: > > Alternatively, you might try to import vg with > lickproto=lock_nolock, then do pvmove. > I'll try that, but with doing so, all LVs from the VG would have to be unmounted, and that's a problem for now. I think that I juste might be doing a new LV on the pv I want to move the data to and just rename the LV, it might be simpler for that situation... I'll ask GSS to see what would be the procedure in a clustered environement... From ekuric at redhat.com Thu Dec 13 14:41:30 2012 From: ekuric at redhat.com (Elvir Kuric) Date: Thu, 13 Dec 2012 15:41:30 +0100 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <50C9E757.4080005@cybercat.ca> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> <50C94177.8070706@cybercat.ca> <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> <50C9E757.4080005@cybercat.ca> Message-ID: <50C9E91A.1080807@redhat.com> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 12/13/2012 03:33 PM, Nicolas Ross wrote: >> Cant you just do a "vgchange -aey" to put the vg into exclusive mode? >> Thats what used to be the way to do it. Its been a while though. >> >> On Dec 12, 2012 8:30 PM, "Yu" > > wrote: >> >> Alternatively, you might try to import vg with >> lickproto=lock_nolock, then do pvmove. >> > > I'll try that, but with doing so, all LVs from the VG would have to be > unmounted, and that's a problem for now. I think that I juste might be > doing a new LV on the pv I want to move the data to and just rename the > LV, it might be simpler for that situation... > > I'll ask GSS to see what would be the procedure in a clustered > environement... > You will need to install 'cmirror' package(s),and start cmirror service on all cluster nodes # service cmirror start After that pvmove should work Here is some docs : - -> https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/5/html-single/Logical_Volume_Manager_Administration/index.html#mirvol_create_ex - -> https://access.redhat.com/knowledge/solutions/279583 ( eventual error if cmirror is not running ) ... but as you said, can you open case with GSS and get dedicated resource for this question Thank you Kind regards, - -- Elvir Kuric,TSE / Red Hat / GSS EMEA / -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) Comment: Using GnuPG with undefined - http://www.enigmail.net/ iF4EAREIAAYFAlDJ6RQACgkQ8YYZ36KGw0PkHAD/aJ2/JPJrlWX4q/AbmVMImWKx QgYOQumm7NGJqZb8PaoA/0Lgw2FY9DLBjK8fO4JYDmP/QggfSHq+tIXnmB4VfW2O =f+k9 -----END PGP SIGNATURE----- From rossnick-lists at cybercat.ca Thu Dec 13 16:18:55 2012 From: rossnick-lists at cybercat.ca (Nicolas Ross) Date: Thu, 13 Dec 2012 11:18:55 -0500 Subject: [Linux-cluster] Moving Physical extents from one PV to another in a clustered environement. In-Reply-To: <50C9E91A.1080807@redhat.com> References: <50C8C89F.9080200@cybercat.ca> <20121212183937.GI14097@squishy.elizium.za.net> <50C8DB71.7010009@cybercat.ca> <20121212202905.GJ14097@squishy.elizium.za.net> <50C94177.8070706@cybercat.ca> <00386E1F-FEC6-4CD2-8BB8-8C61A48E17DE@gmail.com> <50C9E757.4080005@cybercat.ca> <50C9E91A.1080807@redhat.com> Message-ID: <50C9FFEF.6030303@cybercat.ca> > You will need to install 'cmirror' package(s),and start cmirror service > on all cluster nodes > > # service cmirror start > > After that pvmove should work No it didn't. I posted in a previous email what it did, It complains that it cannot lock the vg. > https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/5/html-single/Logical_Volume_Manager_Administration/index.html#mirvol_create_ex > > -> https://access.redhat.com/knowledge/solutions/279583 ( eventual > error if cmirror is not running ) ... > > but as you said, can you open case with GSS and get dedicated resource > for this question The LVs are already created, so to do the operation, I would have to do a lvconvert on the LV to add mirror first ? I'm not familiar at all with lv mirror... From lists at alteeve.ca Mon Dec 17 05:41:30 2012 From: lists at alteeve.ca (Digimer) Date: Mon, 17 Dec 2012 00:41:30 -0500 Subject: [Linux-cluster] Adding VMs to cluster.conf with ccs (or something CLI) Message-ID: <50CEB08A.7010901@alteeve.ca> Hi all, Is there a way to add and remove VM services to cluster.conf/rgmanager (rhcs3 / rhel 6.3) without directly editing the cluster.conf file? I've got a custom application that can create VMs, but I am very nervous about going in and editing -> validating -> pushing out cluster.conf directly. If there is an existing tool, I'd feel much safer using it instead. Thanks! -- Digimer Papers and Projects: https://alteeve.ca/w/ What if the cure for cancer is trapped in the mind of a person without access to education? From epretorious at yahoo.com Mon Dec 17 07:09:44 2012 From: epretorious at yahoo.com (Eric) Date: Sun, 16 Dec 2012 23:09:44 -0800 (PST) Subject: [Linux-cluster] HA iSCSI+DRBD Message-ID: <1355728184.37882.YahooMailNeo@web126002.mail.ne1.yahoo.com> I'd like to share/export DRBD volumes using iSCSI(using a virtual IP address) and use Pacemaker/Corosync to make the iSCSI service highly available. How should I go about coordinating the promotion/demotion of the DRBD resources and the transfer of the virtual IP address (between the two iSCSI targets)? Eric Pretorious Truckee, CA -------------- next part -------------- An HTML attachment was scrubbed... URL: From epretorious at yahoo.com Mon Dec 17 20:51:55 2012 From: epretorious at yahoo.com (Eric) Date: Mon, 17 Dec 2012 12:51:55 -0800 (PST) Subject: [Linux-cluster] HA iSCSI+DRBD In-Reply-To: <1355728184.37882.YahooMailNeo@web126002.mail.ne1.yahoo.com> References: <1355728184.37882.YahooMailNeo@web126002.mail.ne1.yahoo.com> Message-ID: <1355777515.83580.YahooMailNeo@web126004.mail.ne1.yahoo.com> What I should have said was... I've configured a two-node DRBD cluster with a handful of resources/volumes. Each of the volumes is exported/shared from the DRBD primary node using iscsitarget (iSCSI Enterprise Target - IET). The secondary node is not exported/shared using iSCSI - It's only a duplicate of the primary node. How can I use Corosync+Pacemaker to make the iSCSI service highly available using the DRBD secondary node? Eric Pretorious Truckee, CA >________________________________ > From: Eric >To: linux clustering >Sent: Sunday, December 16, 2012 11:09 PM >Subject: [Linux-cluster] HA iSCSI+DRBD > > >I'd like to share/export DRBD volumes using iSCSI(using a virtual IP address) and use Pacemaker/Corosync to make the iSCSI service highly available. > > >How should I go about coordinating the promotion/demotion of the DRBD resources and the transfer of the virtual IP address (between the two iSCSI targets)? > > >Eric Pretorious >Truckee, CA > >-- >Linux-cluster mailing list >Linux-cluster at redhat.com >https://www.redhat.com/mailman/listinfo/linux-cluster > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashish.ext123 at gmail.com Tue Dec 18 05:37:56 2012 From: ashish.ext123 at gmail.com (Ashish G) Date: Tue, 18 Dec 2012 11:07:56 +0530 Subject: [Linux-cluster] CCSD ipv6 bind failures In-Reply-To: References: Message-ID: hi, If anyone knows answers of following questions please let me know. Regards, Ashish On Wed, Dec 5, 2012 at 7:22 PM, Ashish G wrote: > hi Experts, > I has few question on ccsd: > 1. what is the purpose of ccsd listening on ipv4 and ipv6 addresses as > follows in my 2 node HA setup? We do not use IPv6 in our setup. > > netstat -antp |grep ccsd > > tcp 0 0 0.0.0.0:50008 0.0.0.0:* > LISTEN 18170/ccsd > > *tcp 0 0 ::1:50006 > :::* LISTEN 18170/ccsd* > > > ** > > *CCSD is invoked without any arguments: > * > > *root 18170 1 0 14:12 ? 00:00:00 /sbin/ccsd* > > 2. Is it not sufficient to use IPv4 only by CCSD? > > 3. Sometimes on node reboot we see messages as follows: > > "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL > pointer: Address already in use" > Does anyone know when is this problem seen? > > 4. If I invoke CCSD with "-4" option, will I get rid of the above message? > since ipv6 code path will not hit. > > *root 461 1 0 14:28 ? 00:00:00 /sbin/ccsd -4* > > netstat -antp |grep ccsd > > *tcp 0 0 127.0.0.1:50006 0.0.0.0:* > LISTEN 461/ccsd* > > tcp 0 0 0.0.0.0:50008 0.0.0.0:* > LISTEN 461/ccsd > > > Regards, > Ashish > -------------- next part -------------- An HTML attachment was scrubbed... URL: From fdinitto at redhat.com Tue Dec 18 07:17:53 2012 From: fdinitto at redhat.com (Fabio M. Di Nitto) Date: Tue, 18 Dec 2012 08:17:53 +0100 Subject: [Linux-cluster] CCSD ipv6 bind failures In-Reply-To: References: Message-ID: <50D018A1.2090405@redhat.com> On 12/5/2012 2:52 PM, Ashish G wrote: > hi Experts, > I has few question on ccsd: > 1. what is the purpose of ccsd listening on ipv4 and ipv6 addresses as > follows in my 2 node HA setup? We do not use IPv6 in our setup. > > netstat -antp |grep ccsd > > tcp 0 0 0.0.0.0:50008 > 0.0.0.0:* LISTEN 18170/ccsd > > *tcp 0 0 ::1:50006 > :::* LISTEN 18170/ccsd* > > > ** > > *CCSD is invoked without any arguments: > * > > *root 18170 1 0 14:12 ? 00:00:00 /sbin/ccsd* > RFCs now mandates software to support both IPv6 and IPv4 and v6 has to be preferred when available. > > 2. Is it not sufficient to use IPv4 only by CCSD? For operational purposes yes, one protocol is enough, but we need to support both. > > 3. Sometimes on node reboot we see messages as follows: > > "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL > pointer: Address already in use" > > Does anyone know when is this problem seen? No, it?s probably harmless. Does it create any runtime problem? I guess that the kernel ipv6 module is not loaded and bind fails. > > 4. If I invoke CCSD with "-4" option, will I get rid of the above > message? since ipv6 code path will not hit. > > *root 461 1 0 14:28 ? 00:00:00 /sbin/ccsd -4* > > netstat -antp |grep ccsd > > *tcp 0 0 127.0.0.1:50006 > 0.0.0.0:* LISTEN 461/ccsd* > > tcp 0 0 0.0.0.0:50008 > 0.0.0.0:* LISTEN 461/ccsd > Just try it, if you are not using v6, specifying -4 should be enough. Fabio From ashish.ext123 at gmail.com Tue Dec 18 12:33:33 2012 From: ashish.ext123 at gmail.com (Ashish G) Date: Tue, 18 Dec 2012 18:03:33 +0530 Subject: [Linux-cluster] CCSD ipv6 bind failures In-Reply-To: <50D018A1.2090405@redhat.com> References: <50D018A1.2090405@redhat.com> Message-ID: hi Fabio, Thanks for your replies. > "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL > pointer: Address already in use" This issue is not seen always. We see this problem when our application is run for the first time after it is installed. cman start fails as a result of these logs since ccsd fails and it keeps retrying. Rebooting the node helped us get rid of the logs. Its possible that ipv6 module did not load when binding was attempted. But if you look at the log, it says binding failed for "backend" port (50007). Under normal running of system, netstat does not list this port number against ccsd. Instead frontend (50006) and cluster_base_port (50008) are listed. Any idea why backend port binding is attempted and that too only when our application is freshly installed and run? Regards, Ashish On Tue, Dec 18, 2012 at 12:47 PM, Fabio M. Di Nitto wrote: > On 12/5/2012 2:52 PM, Ashish G wrote: > > hi Experts, > > I has few question on ccsd: > > 1. what is the purpose of ccsd listening on ipv4 and ipv6 addresses as > > follows in my 2 node HA setup? We do not use IPv6 in our setup. > > > > netstat -antp |grep ccsd > > > > tcp 0 0 0.0.0.0:50008 > > 0.0.0.0:* LISTEN 18170/ccsd > > > > *tcp 0 0 ::1:50006 > > :::* LISTEN 18170/ccsd* > > > > > > ** > > > > *CCSD is invoked without any arguments: > > * > > > > *root 18170 1 0 14:12 ? 00:00:00 /sbin/ccsd* > > > > RFCs now mandates software to support both IPv6 and IPv4 and v6 has to > be preferred when available. > > > > > 2. Is it not sufficient to use IPv4 only by CCSD? > > For operational purposes yes, one protocol is enough, but we need to > support both. > > > > > 3. Sometimes on node reboot we see messages as follows: > > > > "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL > > pointer: Address already in use" > > > > Does anyone know when is this problem seen? > > No, it?s probably harmless. Does it create any runtime problem? I guess > that the kernel ipv6 module is not loaded and bind fails. > > > > > 4. If I invoke CCSD with "-4" option, will I get rid of the above > > message? since ipv6 code path will not hit. > > > > *root 461 1 0 14:28 ? 00:00:00 /sbin/ccsd -4* > > > > netstat -antp |grep ccsd > > > > *tcp 0 0 127.0.0.1:50006 > > 0.0.0.0:* LISTEN 461/ccsd* > > > > tcp 0 0 0.0.0.0:50008 > > 0.0.0.0:* LISTEN 461/ccsd > > > > Just try it, if you are not using v6, specifying -4 should be enough. > > Fabio > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cfeist at redhat.com Tue Dec 18 21:07:42 2012 From: cfeist at redhat.com (Chris Feist) Date: Tue, 18 Dec 2012 15:07:42 -0600 Subject: [Linux-cluster] Adding VMs to cluster.conf with ccs (or something CLI) In-Reply-To: <50CEB08A.7010901@alteeve.ca> References: <50CEB08A.7010901@alteeve.ca> Message-ID: <50D0DB1E.6020005@redhat.com> On 12/16/12 23:41, Digimer wrote: > Hi all, > > Is there a way to add and remove VM services to cluster.conf/rgmanager > (rhcs3 / rhel 6.3) without directly editing the cluster.conf file? > > I've got a custom application that can create VMs, but I am very > nervous about going in and editing -> validating -> pushing out > cluster.conf directly. If there is an existing tool, I'd feel much safer > using it instead. Yes, you can use --addvm & --rmvm with ccs version 0.16.2-37 and newer (Red Hat Enterprise Linux 6.2 and later). Thanks! Chris ps. I remember talking about this on #linux-cluster, but replying to the list so everyoen can see it. > > Thanks! > From fdinitto at redhat.com Thu Dec 20 07:29:23 2012 From: fdinitto at redhat.com (Fabio M. Di Nitto) Date: Thu, 20 Dec 2012 08:29:23 +0100 Subject: [Linux-cluster] CCSD ipv6 bind failures In-Reply-To: References: <50D018A1.2090405@redhat.com> Message-ID: <50D2BE53.4050107@redhat.com> On 12/18/2012 1:33 PM, Ashish G wrote: > hi Fabio, > Thanks for your replies. > >> "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL >> pointer: Address already in use" > This issue is not seen always. We see this problem when our application > is run for the first time after it is installed. cman start fails as a > result of these logs since ccsd fails and it keeps retrying. Rebooting > the node helped us get rid of the logs. > Its possible that ipv6 module did not load when binding was attempted. > But if you look at the log, it says binding failed for "backend" port > (50007). Under normal running of system, netstat does not list this port > number against ccsd. Instead frontend (50006) and cluster_base_port > (50008) are listed. Any idea why backend port binding is attempted and > that too only when our application is freshly installed and run? No, I don?t know. It?s possible that something is using port 50007 for other reasons and clearly the bind fails. Or maybe an interface is not up. Hard to say without sosreports from a bad and a good boot. If you are a RH customer, please file a ticket with GSS so that we can access the data required to perform debugging. Fabio > > Regards, > Ashish > > On Tue, Dec 18, 2012 at 12:47 PM, Fabio M. Di Nitto > wrote: > > On 12/5/2012 2:52 PM, Ashish G wrote: > > hi Experts, > > I has few question on ccsd: > > 1. what is the purpose of ccsd listening on ipv4 and ipv6 addresses as > > follows in my 2 node HA setup? We do not use IPv6 in our setup. > > > > netstat -antp |grep ccsd > > > > tcp 0 0 0.0.0.0:50008 > > > 0.0.0.0:* LISTEN 18170/ccsd > > > > *tcp 0 0 ::1:50006 > > :::* LISTEN 18170/ccsd* > > > > > > ** > > > > *CCSD is invoked without any arguments: > > * > > > > *root 18170 1 0 14:12 ? 00:00:00 /sbin/ccsd* > > > > RFCs now mandates software to support both IPv6 and IPv4 and v6 has to > be preferred when available. > > > > > 2. Is it not sufficient to use IPv4 only by CCSD? > > For operational purposes yes, one protocol is enough, but we need to > support both. > > > > > 3. Sometimes on node reboot we see messages as follows: > > > > "Unable to bind to backend ipv6 socket, but inet_ntop returned NULL > > pointer: Address already in use" > > > > Does anyone know when is this problem seen? > > No, it?s probably harmless. Does it create any runtime problem? I guess > that the kernel ipv6 module is not loaded and bind fails. > > > > > 4. If I invoke CCSD with "-4" option, will I get rid of the above > > message? since ipv6 code path will not hit. > > > > *root 461 1 0 14:28 ? 00:00:00 /sbin/ccsd -4* > > > > netstat -antp |grep ccsd > > > > *tcp 0 0 127.0.0.1:50006 > > > 0.0.0.0:* LISTEN 461/ccsd* > > > > tcp 0 0 0.0.0.0:50008 > > > 0.0.0.0:* LISTEN 461/ccsd > > > > Just try it, if you are not using v6, specifying -4 should be enough. > > Fabio > > -- > Linux-cluster mailing list > Linux-cluster at redhat.com > https://www.redhat.com/mailman/listinfo/linux-cluster > > > > From a.holway at syseleven.de Sat Dec 29 02:49:04 2012 From: a.holway at syseleven.de (Andrew Holway) Date: Sat, 29 Dec 2012 03:49:04 +0100 Subject: [Linux-cluster] Poor LVM performance. Message-ID: Hi, I have been asking around about this for a while. I got the same results with CLVM with an iSCSI box I had on loan. I have been doing some testing with KVM and Virtuozzo(containers based virtualisation) and various storage devices and have some results I would like some help analyzing. I have a nice big ZFS box from Oracle (Yes, evil but Solaris NFS is amazing). I have 10G and IB connecting these to my cluster. My cluster is four HP servers (E5-2670 & 144GB ram) with a RAID10 of 600k SAS drives. Please open these pictures side by side. https://dl.dropbox.com/u/98200887/Screen%20Shot%202012-12-04%20at%202.50.33%20PM.png https://dl.dropbox.com/u/98200887/Screen%20Shot%202012-12-04%20at%203.18.03%20PM.png You will notice that using KVM/LVM on the local RAID10 (and CLVM on iSCSI) completely destroys performance whereas the container based virtualisation stuff is awesome and as fast as the NFS. 4,8,12,16...VMs relates to the aggregate performance of the benchmark in that number of VMs. 4 = 1 VM on each node, 8 = 2 VM on each node. TPCC warehouses is the number of tpcc warehouses that the benchmark used. 1 warehouse is about 150MB so 10 warehouses would mean about 1.5GB of data being held in the innodb pool. Why does LVM performance suck so hard compared to a single filesystem approach. What am I doing wrong? Thanks, Andrew