Re: [ovirt-users] ovirt - unsupported GFS config ??

This is a multi-part message in MIME format. --------------060409020502060901040200 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit On 03/27/2016 10:24 PM, paf1@email.cz wrote:
What's the recommended shard size for databases ( espacially Oracle ) , I'm afraid that 512M is too large. I found that someones using about 16MB , but it generate a lot of files for healing if volumes splitted. (eg. for 500GB DB in worse case )
Would the database be running within the guest VM? Did you run into any specific issue with 512M shard size? What we have noticed is that with smaller shard sizes , like 4MB, the number of entries in the .shard directory is too high, and hence affects the performance when it comes to healing entries in directory. The 256M/512M shard size is a good balance between the number of entries created and data size to heal.
Pa.
On 27.3.2016 16:57, Sahina Bose wrote:
Stripe is not supported.
What you need to do instead is turn on sharding for the volume.
So:
gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS force
gluster volume set 12HP12-S2R3A1P2 features.shard on gluster volume set 12HP12-S2R3A1P2 features.shard-block-size 512MB
If you want to utilize the additional nodes as well, you can change this to a distributed replicate volume - instead of the volume creation in step above , use below
gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force
On 03/24/2016 07:49 PM, paf1@email.cz wrote:
Hello, I tried create stripe 2 replica 3 arbiter1 gluster volume for testing. So , glusterFS such type from commandline was successfull, but domain creation looks to be unsupported. with oVirt message "Error while executing action AddGlusterFsStorageDomain: Storage Domain target is unsupported". Cam U tell me if is it error or really unsuported ??
exam: gluster volume create 12HP12-S2R3A1P2 stripe 2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force
RHEL 7-2.1511 vdsm - vdsm-4.17.23-1.el7 gluster - glusterfs-3.7.9-1.el7 ovirt - 3.5.6.2-1
regs.Pavel
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
--------------060409020502060901040200 Content-Type: text/html; charset=windows-1252 Content-Transfer-Encoding: 8bit <html> <head> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> </head> <body text="#000000" bgcolor="#FFFFFF"> <br> <br> <br> <div class="moz-cite-prefix">On 03/27/2016 10:24 PM, <a class="moz-txt-link-abbreviated" href="mailto:paf1@email.cz">paf1@email.cz</a> wrote:<br> </div> <blockquote cite="mid:56F81046.3070200@email.cz" type="cite"> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> What's the recommended shard size for databases ( espacially Oracle ) , I'm afraid that 512M is too large.<br> I found that someones using about 16MB , but it generate a lot of files for healing if volumes splitted. (eg. for 500GB DB in worse case )<br> </blockquote> <br> Would the database be running within the guest VM?<br> <br> Did you run into any specific issue with 512M shard size?<br> What we have noticed is that with smaller shard sizes , like 4MB, the number of entries in the .shard directory is too high, and hence affects the performance when it comes to healing entries in directory. The 256M/512M shard size is a good balance between the number of entries created and data size to heal.<br> <br> <blockquote cite="mid:56F81046.3070200@email.cz" type="cite"> <br> Pa.<br> <br> <div class="moz-cite-prefix">On 27.3.2016 16:57, Sahina Bose wrote:<br> </div> <blockquote cite="mid:56F7F4C0.2020408@redhat.com" type="cite"> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> Stripe is not supported.<br> <br> What you need to do instead is turn on sharding for the volume.<br> <br> So:<br> <br> gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS force<br> <br> gluster volume set 12HP12-S2R3A1P2 features.shard on<br> gluster volume set 12HP12-S2R3A1P2 features.shard-block-size 512MB<br> <br> If you want to utilize the additional nodes as well, you can change this to a distributed replicate volume - instead of the volume creation in step above , use below<br> <br> gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force<br> <br> <br> <div class="moz-cite-prefix">On 03/24/2016 07:49 PM, <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:paf1@email.cz"><a class="moz-txt-link-abbreviated" href="mailto:paf1@email.cz">paf1@email.cz</a></a> wrote:<br> </div> <blockquote cite="mid:56F3F75D.8060201@email.cz" type="cite"> <meta http-equiv="content-type" content="text/html; charset=windows-1252"> Hello, <br> I tried create stripe 2 replica 3 arbiter1 gluster volume for testing.<br> So , glusterFS such type from commandline was successfull, but domain creation looks to be unsupported. with oVirt message "Error while executing action AddGlusterFsStorageDomain: Storage Domain target is unsupported".<br> Cam U tell me if is it error or really unsuported ??<br> <br> exam: <br> gluster volume create 12HP12-S2R3A1P2 stripe 2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force<br> <br> <br> RHEL 7-2.1511<br> vdsm - vdsm-4.17.23-1.el7<br> gluster - glusterfs-3.7.9-1.el7<br> ovirt - 3.5.6.2-1<br> <br> regs.Pavel<br> <br> <fieldset class="mimeAttachmentHeader"></fieldset> <br> <pre wrap="">_______________________________________________ Users mailing list <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:Users@ovirt.org">Users@ovirt.org</a> <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://lists.ovirt.org/mailman/listinfo/users">http://lists.ovirt.org/mailman/listinfo/users</a> </pre> </blockquote> <br> </blockquote> <br> </blockquote> <br> </body> </html> --------------060409020502060901040200--

This is a multi-part message in MIME format. --------------010809000104040808080300 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit - Yes, database will run in VM , but with RMAN online backups. So VM takes over 1,4TB . This VM wil run 5-6x :o( - I have no special issues with 512MB shard size, looks OK but not tested for such databases yet ( Waiting for 3.7.10 gluster (dupl.healing fix)) regs. Pavel On 28.3.2016 08:30, Sahina Bose wrote:
On 03/27/2016 10:24 PM, paf1@email.cz wrote:
What's the recommended shard size for databases ( espacially Oracle ) , I'm afraid that 512M is too large. I found that someones using about 16MB , but it generate a lot of files for healing if volumes splitted. (eg. for 500GB DB in worse case )
Would the database be running within the guest VM?
Did you run into any specific issue with 512M shard size? What we have noticed is that with smaller shard sizes , like 4MB, the number of entries in the .shard directory is too high, and hence affects the performance when it comes to healing entries in directory. The 256M/512M shard size is a good balance between the number of entries created and data size to heal.
Pa.
On 27.3.2016 16:57, Sahina Bose wrote:
Stripe is not supported.
What you need to do instead is turn on sharding for the volume.
So:
gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS force
gluster volume set 12HP12-S2R3A1P2 features.shard on gluster volume set 12HP12-S2R3A1P2 features.shard-block-size 512MB
If you want to utilize the additional nodes as well, you can change this to a distributed replicate volume - instead of the volume creation in step above , use below
gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force
On 03/24/2016 07:49 PM, paf1@email.cz wrote:
Hello, I tried create stripe 2 replica 3 arbiter1 gluster volume for testing. So , glusterFS such type from commandline was successfull, but domain creation looks to be unsupported. with oVirt message "Error while executing action AddGlusterFsStorageDomain: Storage Domain target is unsupported". Cam U tell me if is it error or really unsuported ??
exam: gluster volume create 12HP12-S2R3A1P2 stripe 2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force
RHEL 7-2.1511 vdsm - vdsm-4.17.23-1.el7 gluster - glusterfs-3.7.9-1.el7 ovirt - 3.5.6.2-1
regs.Pavel
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
--------------010809000104040808080300 Content-Type: text/html; charset=windows-1252 Content-Transfer-Encoding: 8bit <html> <head> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> </head> <body text="#000066" bgcolor="#FFFFFF"> - Yes, database will run in VM , but with RMAN online backups. So VM takes over 1,4TB . This VM wil run 5-6x :o(<br> - I have no special issues with 512MB shard size, looks OK but not tested for such databases yet ( Waiting for 3.7.10 gluster (dupl.healing fix))<br> <br> regs.<br> Pavel<br> <br> <div class="moz-cite-prefix">On 28.3.2016 08:30, Sahina Bose wrote:<br> </div> <blockquote cite="mid:56F8CF8A.7040306@redhat.com" type="cite"> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> <br> <br> <br> <div class="moz-cite-prefix">On 03/27/2016 10:24 PM, <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:paf1@email.cz">paf1@email.cz</a> wrote:<br> </div> <blockquote cite="mid:56F81046.3070200@email.cz" type="cite"> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> What's the recommended shard size for databases ( espacially Oracle ) , I'm afraid that 512M is too large.<br> I found that someones using about 16MB , but it generate a lot of files for healing if volumes splitted. (eg. for 500GB DB in worse case )<br> </blockquote> <br> Would the database be running within the guest VM?<br> <br> Did you run into any specific issue with 512M shard size?<br> What we have noticed is that with smaller shard sizes , like 4MB, the number of entries in the .shard directory is too high, and hence affects the performance when it comes to healing entries in directory. The 256M/512M shard size is a good balance between the number of entries created and data size to heal.<br> <br> <blockquote cite="mid:56F81046.3070200@email.cz" type="cite"> <br> Pa.<br> <br> <div class="moz-cite-prefix">On 27.3.2016 16:57, Sahina Bose wrote:<br> </div> <blockquote cite="mid:56F7F4C0.2020408@redhat.com" type="cite"> <meta content="text/html; charset=windows-1252" http-equiv="Content-Type"> Stripe is not supported.<br> <br> What you need to do instead is turn on sharding for the volume.<br> <br> So:<br> <br> gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS force<br> <br> gluster volume set 12HP12-S2R3A1P2 features.shard on<br> gluster volume set 12HP12-S2R3A1P2 features.shard-block-size 512MB<br> <br> If you want to utilize the additional nodes as well, you can change this to a distributed replicate volume - instead of the volume creation in step above , use below<br> <br> gluster volume create 12HP12-S2R3A1P2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force<br> <br> <br> <div class="moz-cite-prefix">On 03/24/2016 07:49 PM, <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:paf1@email.cz">paf1@email.cz</a> wrote:<br> </div> <blockquote cite="mid:56F3F75D.8060201@email.cz" type="cite"> <meta http-equiv="content-type" content="text/html; charset=windows-1252"> Hello, <br> I tried create stripe 2 replica 3 arbiter1 gluster volume for testing.<br> So , glusterFS such type from commandline was successfull, but domain creation looks to be unsupported. with oVirt message "Error while executing action AddGlusterFsStorageDomain: Storage Domain target is unsupported".<br> Cam U tell me if is it error or really unsuported ??<br> <br> exam: <br> gluster volume create 12HP12-S2R3A1P2 stripe 2 replica 3 arbiter 1 1hp1:/STORAGES/P2/GFS 1hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-1/GFS 2hp1:/STORAGES/P2/GFS 2hp2:/STORAGES/P2/GFS kvmarbiter:/STORAGES/P2-2/GFS force<br> <br> <br> RHEL 7-2.1511<br> vdsm - vdsm-4.17.23-1.el7<br> gluster - glusterfs-3.7.9-1.el7<br> ovirt - 3.5.6.2-1<br> <br> regs.Pavel<br> <br> <fieldset class="mimeAttachmentHeader"></fieldset> <br> <pre wrap="">_______________________________________________ Users mailing list <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:Users@ovirt.org">Users@ovirt.org</a> <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://lists.ovirt.org/mailman/listinfo/users">http://lists.ovirt.org/mailman/listinfo/users</a> </pre> </blockquote> <br> </blockquote> <br> </blockquote> <br> </blockquote> <br> </body> </html> --------------010809000104040808080300--
participants (2)
-
paf1@email.cz
-
Sahina Bose