<html><body><div style="font-family: arial,helvetica,sans-serif; font-size: 10pt; color: #000000">That radical of a performance slowdown is often a network problem, especially if this is a new hardware setup that hasn't been shaken out yet. Can you crank up a copy of Wireshark and see if there is something funky going on at that level (like lots of timeouts)?<br><br><hr id="zwchr"><div style="color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt;"><b>From: </b>"Joel Young" <jdy@cryregarder.com><br><b>To: </b>"Nux!" <nux@li.nux.ro><br><b>Cc: </b>"gluster-users" <gluster-users@gluster.org><br><b>Sent: </b>Friday, October 11, 2013 10:47:11 AM<br><b>Subject: </b>Re: [Gluster-users] Tuning for Home Directories<br><br>Yes. No appreciable improvement.<br><br>Thanks!<br><br>Joel<br><br>On Fri, Oct 11, 2013 at 7:44 AM, Nux! <nux@li.nux.ro> wrote:<br>> On 11.10.2013 15:33, Joel Young wrote:<br>>><br>>> Re home directory tuning...<br>>><br>>> Any suggestions? We are still getting extraordinary slowness in using<br>>> gluster for home directories. SVN checkouts to the volume are very<br>>> slow, configure scripts are very slow.<br>>><br>>> The cluster hardware is overspecced. Each of the four nodes are<br>>> connected via 10GB ethernet, have 24-64 cores, and from 128GB to<br>>> 512GB.<br>>><br>>> Thanks!<br>>><br>>> On Wed, Sep 18, 2013 at 10:44 AM, Joel Young <jdy@cryregarder.com> wrote:<br>>>><br>>>> I've got a HPC type set up with four nodes and a<br>>>> distributed-replicated configuration. Performing developer type<br>>>> actions is disfunctionally slow on the volume. Below is the<br>>>> configuration:<br>>>><br>>>> Volume Name: home<br>>>> Type: Distributed-Replicate<br>>>> Volume ID: 83fa39a6-6e68-4e1c-8fae-3c3e30b1bd66<br>>>> Status: Started<br>>>> Number of Bricks: 2 x 2 = 4<br>>>> Transport-type: tcp<br>>>> Bricks:<br>>>> Brick1: ir0:/lhome/gluster_home<br>>>> Brick2: ir1:/lhome/gluster_home<br>>>> Brick3: ir2:/lhome/gluster_home<br>>>> Brick4: ir3:/raid/gluster_home<br>>>> Options Reconfigured:<br>>>> server.statedump-path: /tmp<br>>>> performance.cache-size: 512MB<br>>>> performance.client-io-threads: on<br>>>><br>>>> For example, configure runs take tens of minutes instead of fractions<br>>>> of minutes.<br>>>><br>>>> I've tried setting lookup-unhashed to off or no without a significant<br>>>> performance change.<br>>>><br>>>> Does anyone have any performance tunings or suggestions for making<br>>>> gluster based home directories usable for developers?<br>>>><br>>>> Also, the documentation is unclear as to if lookup-unhashed should be<br>>>> set to yes or on, vs no or off. Some things indicate one way, some<br>>>> the other. boolean options should no allow multiple values to be set.<br>><br>><br>> Have you tried mounting the volume over NFS and see how that goes?<br>><br>> Lucian<br>><br>> --<br>> Sent from the Delta quadrant using Borg technology!<br>><br>> Nux!<br>> www.nux.ro<br>_______________________________________________<br>Gluster-users mailing list<br>Gluster-users@gluster.org<br>http://supercolony.gluster.org/mailman/listinfo/gluster-users<br></div><br/></div></body></html>