Well, it took us a long time to actually deploy a working 2.0 server. The default method of using a bridge broke the all nfs-root setup we had(our cpu nodes are diskless; makes them easier to maintain). For a while, we had been running a busy mail server thru xen(over nfs). It had problems with both 1.0 and 1.2(had to restart the instance every week or so). When we got 2.0 working at the start of this week, the first instance we moved(we have 17 or so) was this mail server. The next day, the server started getting flooded with spam(when I finally discovered the issue, it had over 250k mails in it''s queue). In the process of fixing that, I had a sub-optimal config, that generated a load of 1000(ran out of file descriptors, and the kernel was spewing that message constantly to the console). While the xen instance itself was rather unresponsive, logging into dom0 was rather speedy(if you ignore that dom0 was exporting a swap device over nfs to the instance, which was pushing it''s cache out). So far, we are extremely happy with the stability of 2.0. Congrats. ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
> So far, we are extremely happy with the stability of 2.0. Congrats.That''s good to hear. Do you still use NFS root? Do you use 2.4 or 2.6 kernels?> Well, it took us a long time to actually deploy a working 2.0 server. The > default method of using a bridge broke the all nfs-root setup we had(our cpu > nodes are diskless; makes them easier to maintain).I want to get example scripts for a routed (as opposed to bridged setup) checked into the tree... Thansk, Ian ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
On Fri, 7 Jan 2005, Ian Pratt wrote:> > > So far, we are extremely happy with the stability of 2.0. Congrats. > > That''s good to hear. > > Do you still use NFS root? Do you use 2.4 or 2.6 kernels?Yup. Pure nfsroot. pxegrub, nfsroot dom0, nfsroot domU. The only change, is that for 1.0/1.2, we used dhcp on the domU nodes. For 2.0, we use static routing.> > Well, it took us a long time to actually deploy a working 2.0 server. The > > default method of using a bridge broke the all nfs-root setup we had(our cpu > > nodes are diskless; makes them easier to maintain). > > I want to get example scripts for a routed (as opposed to bridged > setup) checked into the tree...I''m not fully satified with the scripts; that abuse parameters for their own nefarious purposes. Maybe when we convert the second machine to 2.0(it''s 1.2 atm). ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel