Re: [Gluster-users] To GlusterFS or not...
Yes, Roman is correct. Also, if you have lots of random IO you're better off with many smaller SAS drives. This is because the greater number of spindles you have the greater your random IO is. This is also why we went with ssd drives because sas drives weren't cutting it on the random io front. Another option you may try is using SAS drives with ZFS compression. Compression will be especially helpful if you're using SATA drives. -Alex On Sep 23, 2014 2:10 PM, Roman rome...@gmail.com wrote: Hi, SAS 7200 RPM disks are not that small size at all (same as SATA basically). If I remember right, the reason of switching to SAS here would be Full Duplex with SAS (you can read and write in the same time to them) instead of Half Duplex with SATA disks (read or write per one moment only). 2014-09-23 9:02 GMT+03:00 Chris Knipe sav...@savage.za.org: Hi, SSD has been considered but is not an option due to cost. SAS has been considered but is not a option due to the relatively small sizes of the drives. We are *rapidly* growing towards a PB of actual online storage. We are exploring raid controllers with onboard SSD cache which may help. On Tue, Sep 23, 2014 at 7:59 AM, Roman rome...@gmail.com wrote: Hi, just a question ... Would SAS disks be better in situation with lots of seek times using GlusterFS? 2014-09-22 23:03 GMT+03:00 Jeff Darcy jda...@redhat.com: The biggest issue that we are having, is that we are talking about -billions- of small (max 5MB) files. Seek times are killing us completely from what we can make out. (OS, HW/RAID has been tweaked to kingdom come and back). This is probably the key point. It's unlikely that seek times are going to get better with GlusterFS, unless it's because the new servers have more memory and disks, but if that's the case then you might as well just deploy more memory and disks in your existing scheme. On top of that, using any distributed file system is likely to mean more network round trips, to maintain consistency. There would be a benefit from letting GlusterFS handle the distribution (and redistribution) of files automatically instead of having to do your own sharding, but that's not the same as a performance benefit. I’m not yet too clued up on all the GlusterFS naming, but essentially if we do go the GlusterFS route, we would like to use non replicated storage bricks on all the front-end, as well as back-end servers in order to maximize storage. That's fine, so long as you recognize that recovering from a failed server becomes more of a manual process, but it's probably a moot point in light of the seek-time issue mentioned above. As much as I hate to discourage people from using GlusterFS, it's even worse to have them be disappointed, or for other users with other needs to be so as we spend time trying to fix the unfixable. ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users -- Best regards, Roman. -- Regards, Chris Knipe -- Best regards, Roman. ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] Gluster = RAID 10 over the network?
Hi Ryan, I think if you could provide more info on the storage systems it would help. Things like total drives per raid set and size of each drive. This is a complicated question, but a simple Googling brings up this interesting article: http://wolfcrow.com/blog/which-is-the-best-raid-level-for-video-editing-and-post-production-part-three-number-soup-for-the-soul/ Imho, without knowing any of these details, my personal preference, unless you're running a database is to do multiple raid-1 sets, stripe them with lvm and drop xfs on them. I would like to add that if your storage provider only offers raid-5 or raid-10 it might behoove you to look for another storage provider. :) -Alex On Sep 21, 2014 8:24 PM, Ryan Nix ryan@gmail.com wrote: Hi All, So my boss and I decided to make a good size investment in a Gluster cluster. I'm super excited and I will be taking a Redhat Storage class soon. However, we're debating the hardware configuration we intend to purchase. We agree that each brick/node, and we're buying four, each configured as RAID 10 will help us sleep at night, but to me, it seems like such an unfortunate waste of disk space. Our graduate and PHD students work with lots of video and they filled up our proof-of-concept 4 TB ownCloud/Gluster setup in 2 months. I stumbled upon Howtoforge's Gluster setup guide from two years ago and I'm wondering if this is correct and or still relevant: http://bit.ly/1qkLoVe *This tutorial shows how to combine four single storage servers (running Ubuntu 12.10) to a distributed replicated storage with GlusterFS http://www.gluster.org/. Nodes 1 and 2 (replication1) as well as 3 and 4 (replication2) will mirror each other, and replication1 and replication2 will be combined to one larger storage server (distribution). Basically, this is RAID10 over network. If you lose one server from replication1 and one from replication2, the distributed volume continues to work. The client system (Ubuntu 12.10 as well) will be able to access the storage as if it was a local filesystem* The vendor we have chosen, System 76, offers either RAID 5 or RAID 10 in each server. Does anyone have insights or opinions on this? It would seem to be that RAID 5 would be okay and that some kind drive monitoring (opinions also welcome, please) would be sufficient with the inherent nature of Gluster's Distributed/Replicated setup. RAID 5 at System 76 allows us to max out at 42 TB of useable space. RAID 10 makes it 24 TB useable. I'd love to hear any insights or opinions on this. To me, RAID 5 with Gluster in a distributed replicated setup should be sufficient and help us sleep well each night. :) Thanks in advance! Ryan ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] GlusterFS-3.4.4 RPMs on download.gluster.org
Changelog? On Mon, Jun 16, 2014 at 9:24 PM, Kaleb S. KEITHLEY kkeit...@redhat.com wrote: RPMs for el5-7 (RHEL, CentOS, etc.) and Fedora (19, 20, 21/rawhide), are now available in YUM repos at http://download.gluster.org/pub/gluster/glusterfs/3.4/LATEST There are also RPMs available for Pidora 20, SLES 11sp3 and OpenSuSE 13.1. Debian and Ubuntu DPKGs should also be appearing soon. -- Kaleb ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] GlusterFS-3.4.4 RPMs on download.gluster.org
And, I found it myself: https://github.com/gluster/glusterfs/blob/release-3.4/doc/release-notes/3.4.4.md On Mon, Jun 16, 2014 at 11:13 PM, Alexey Zilber alexeyzil...@gmail.com wrote: Changelog? On Mon, Jun 16, 2014 at 9:24 PM, Kaleb S. KEITHLEY kkeit...@redhat.com wrote: RPMs for el5-7 (RHEL, CentOS, etc.) and Fedora (19, 20, 21/rawhide), are now available in YUM repos at http://download.gluster.org/pub/gluster/glusterfs/3.4/LATEST There are also RPMs available for Pidora 20, SLES 11sp3 and OpenSuSE 13.1. Debian and Ubuntu DPKGs should also be appearing soon. -- Kaleb ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users
[Gluster-users] Cannot create volume on fresh install of 3.4.3 on CentOs 5.10
Hi All, I'm having a horrid time getting gluster to create a volume. Initially, I messed up a path and had the error mentioned here: http://joejulian.name/blog/glusterfs-path-or-a-prefix-of-it-is-already-part-of-a-volume/ I fixed it, restarted gluster on both nodes, and then I just get a straight up failure: # gluster volume create devroot replica 2 transport tcp sfdev1:/data/brick1/devroot sfdev2:/data/brick1/devroot volume create: devroot: failed The only thing that gets created are the extended attributes in /data/brick1/devroot on sfdev1. Here's the cli.log.. not much in there: --- [2014-06-14 05:42:03.519769] W [rpc-transport.c:175:rpc_transport_load] 0-rpc-transport: missing 'option transport-type'. defaulting to socket [2014-06-14 05:42:03.523525] I [socket.c:3480:socket_init] 0-glusterfs: SSL support is NOT enabled [2014-06-14 05:42:03.523580] I [socket.c:3495:socket_init] 0-glusterfs: using system polling thread [2014-06-14 05:42:03.600482] I [cli-cmd-volume.c:392:cli_cmd_volume_create_cbk] 0-cli: Replicate cluster type found. Checking brick order. [2014-06-14 05:42:03.600844] I [cli-cmd-volume.c:304:cli_cmd_check_brick_order] 0-cli: Brick order okay [2014-06-14 05:42:03.668257] I [cli-rpc-ops.c:805:gf_cli_create_volume_cbk] 0-cli: Received resp to create volume [2014-06-14 05:42:03.668365] I [input.c:36:cli_batch] 0-: Exiting with: -1 --- .cmd_log_history shows: [2014-06-14 05:42:03.668051] : volume create devroot replica 2 transport tcp sfdev1:/data/brick1/devroot sfdev2:/data/brick1/devroot : FAILED : etc-glusterfs-glusterd.vol.log seems good too. Any ideas? -Alex ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] Cannot create volume on fresh install of 3.4.3 on CentOs 5.10
:16:30.472989] D [glusterd-utils.c:4980:glusterd_friend_find_by_hostname] 0-management: Friend sfdev2 found.. state: 3 -Alex On Sat, Jun 14, 2014 at 6:32 PM, Vijay Bellur vbel...@redhat.com wrote: On 06/14/2014 11:38 AM, Alexey Zilber wrote: Hi All, I'm having a horrid time getting gluster to create a volume. Initially, I messed up a path and had the error mentioned here: http://joejulian.name/blog/glusterfs-path-or-a-prefix-of- it-is-already-part-of-a-volume/ I fixed it, restarted gluster on both nodes, and then I just get a straight up failure: # gluster volume create devroot replica 2 transport tcp sfdev1:/data/brick1/devroot sfdev2:/data/brick1/devroot volume create: devroot: failed The only thing that gets created are the extended attributes in /data/brick1/devroot on sfdev1. Here's the cli.log.. not much in there: --- [2014-06-14 05:42:03.519769] W [rpc-transport.c:175:rpc_transport_load] 0-rpc-transport: missing 'option transport-type'. defaulting to socket [2014-06-14 05:42:03.523525] I [socket.c:3480:socket_init] 0-glusterfs: SSL support is NOT enabled [2014-06-14 05:42:03.523580] I [socket.c:3495:socket_init] 0-glusterfs: using system polling thread [2014-06-14 05:42:03.600482] I [cli-cmd-volume.c:392:cli_cmd_volume_create_cbk] 0-cli: Replicate cluster type found. Checking brick order. [2014-06-14 05:42:03.600844] I [cli-cmd-volume.c:304:cli_cmd_check_brick_order] 0-cli: Brick order okay [2014-06-14 05:42:03.668257] I [cli-rpc-ops.c:805:gf_cli_create_volume_cbk] 0-cli: Received resp to create volume [2014-06-14 05:42:03.668365] I [input.c:36:cli_batch] 0-: Exiting with: -1 --- .cmd_log_history shows: [2014-06-14 05:42:03.668051]: volume create devroot replica 2 transport tcp sfdev1:/data/brick1/devroot sfdev2:/data/brick1/devroot : FAILED : etc-glusterfs-glusterd.vol.log seems good too. Can you please check again the part relating to the attempt to create volume devroot in glusterd.vol.log? glusterd's log file normally contains information regarding why an operation failed. -Vijay ___ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users