Hi Mark; I have installed several 7000 series systems, some running 100's of VM's. I can help try to help you but to find where exactly the problem is I may need more information.
I can understand that you have no ZIL's. So most probably you are using the 7110 with 250 GB drives. All 7000 series have a module called analytics where you can monitor many components performance. Please start with selecting enable advanced analytics in preferences tab at configuration menu. Please make sure that you are running latest FW release Q1.2.1 2010 @ http://wikis.sun.com/display/FishWorks/Software+Updates Please read all release notes attached from the FW you are running to the FW level you will upgrade to. I understand that you are using iSCSI, if you are running earlier FW's, NFS can increase performance significantly however for recent FW's iSCSI and NFS performance is very close but I'd choose NFS over iSCSI for most installiations. Do so if yu can. Please start monitoring fallowing datasets using analytics. Network transfer broken by interface or device [check if you are stuck by gigabit ethernet etc....] iSCSI IOPS iSCSI IOPS broken down by LUN (to understand which LUN demands more performance, with newer FW's you may find it use full to isolate some LUN's by defining different pools - beware that this may not offer much help if you use Raid 10) iSCSI the iops broken down type iSCSI write iops latency iSCSI latency Arc hit/Miss ratio Arc size Here are my recomendations :(if you can share some screen shots from analytics I may be able to help mre) 1) Convert to Raid10 - this will provide you 4-5x More IOPS on both read and writes. 2) Using analytics, decide if increasing L1 cache may help you. If it can, increase L1 cache 3) Check the IO size using analytics and check it against your lun definations. I suggest that Lun block size should be lower than IO size. 4) Enable wirte caching for a short time and monitor analytics report and if you see much improvement you can invest in SSD's. 5) Enable jumb frames (through out the path) 6) Use multiple interfaces to access data PS: I think that you have asked if you can disable ZIL on 7000 series. The answer is yes and you can decide it at share/lun granularity. PS: We usualy recommend to use writezilla for vmware users but I have seen 7210's running 30-40 vm's without much problem when there were no writezilla, but for sure this depends on the load pattern. Very best regards Mertol Mertol Ozyoney Storage Practice - Sales Manager Sun Microsystems, TR Istanbul TR Phone +902123352200 Mobile +905339310752 Fax +902123352222 Email mertol.ozyo...@sun.com -----Original Message----- From: zfs-discuss-boun...@opensolaris.org [mailto:zfs-discuss-boun...@opensolaris.org] On Behalf Of Mark Sent: Friday, August 27, 2010 10:47 PM To: zfs-discuss@opensolaris.org Subject: Re: [zfs-discuss] VM's on ZFS - 7210 It does, its on a pair of large APC's. Right now we're using NFS for our ESX Servers. The only iSCSI LUN's I have are mounted inside a couple Windows VM's. I'd have to migrate all our VM's to iSCSI, which I'm willing to do if it would help and not cause other issues. So far the 7210 Appliance has been very stable. I like the zilstat script. I emailed a support tech I am working with on another issue to ask if one of the built in Analytics DTrace scripts will get that data. I found one called L2ARC Eligibility: 3235 true, 66 false. This makes it sound like we would benefit from a READZilla, not quite what I had expected... I'm sure I don't know what I'm looking at anyways :) -- This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
