Peter Tribble wrote: > On Sat, Nov 15, 2008 at 10:26 AM, Ian Collins <[EMAIL PROTECTED]> wrote: > >> Anyone who follows this list we have seen a number of issues with >> Solaris 10 and ZFS from me this week. >> >> We deployed Solaris 10 for the usual conservative reasons, support and >> stability. Most of my my ZFS experience has been with SXCE and I've >> seen problems reported and fixed a couple of builds later. The further >> SXCE moves ahead of Solaris 10 ZFS, the longer (and probably more >> difficult) the task of back porting these fixes will become. >> >> So my question is, for production servers (x4540) that are primarily SMB >> (80%) and NFS (20%) file servers, would you deploy SXCE with native >> CIFS support, or Solaris 10/Samba? >> >> I wouldn't hesitate to go with the former, relying on Live Upgrade to >> incorporate fixes rather than patching. Persuading clients may be a >> little harder! >> > > At the present time, I would go with Solaris 10. I don't have a problem > with SXCE as such (although many recent builds have had some issues - > not necessarily with zfs); the problem I do have is with whether SXCE has > a sustainable future. Solaris 10 has years of support left in it, but what > happens once SXCE is scrapped and you can't update any further? > > My main concern is the sate of play now. I've already managed to badly crash Solaris 10 twice with zfs send/receive and we've pretty much ruled Solaris 10 with zfs send/receive out for data replication. I'm going to repeat my testing with SXCE, the zfs receive performance fixes should be in the next spin.
-- Ian. _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
