On Thu, Feb 20, 2003 at 01:45:34PM +0800, Ian C. Sison wrote: > Teka teka, you've dropped XFS completely? I remember you had the best > reviews of XFS a while ago, why all of a sudden the data corruption?
Problems with Gusi, our primary do-it-all server, began to crop up around 2002-09-30, with a message[1] I sent to both the lkml and the linux-xfs list. Things seem to have stabilized a bit until a few days later, with another significant message[2] I sent to the linux-xfs list exclusively after attempts to mount our XFS volumes got stuck on boot. [1] http://marc.theaimsgroup.com/?l=linux-xfs&m=103338806912329&w=2 [2] http://marc.theaimsgroup.com/?l=linux-xfs&m=103369234209733&w=2 I eventually migrated the server to ext3. I can't remember if I did it overnight or in mid-day because of the perennial problems we suddenly started having with XFS (this wasn't a fresh new kernel, BTW, our system had been up and running for 51 days before the problem, and then suddenly we just kept getting these problems during that week). ext3 is slower than XFS, no doubt about that. I run a "poor man's" data backup every other night, and had to adjust the time that the cron job would start because with ext3 the work was spilling onto the next business day which starts at 7:00am. I can sleep much better at night with ext3, though. As soon as I had straightened things out on Gusi using kernel 2.4.19, it kept chugging along until an extended power outage reset the uptime last week, by which time it had been running for 110 days. I took the opportunity to upgrade to 2.4.20, and it has been running for 7 days since then. I am hoping Meralco will allow us to reach higher uptime peaks than our current record[3]. :) [3] http://mrtg.leathercollection.ph/uptime.html > Is is because you're using an unstable patch? And if so, why not revert > to a stable one? Gusi had not been running an unstable patch with XFS. We were using a fairly accepted CVS checkout, with only RML's preempt patches and Randy Dunlop's sysmagic patch to allow me to send SysMagic commands via a /proc interface. Buwaya, my workstation at home, is another story. I had not been able to revert to ext3 together with Gusi, so it still ran XFS until last week. Lately I have been playing around with Con Kolivas's patches, which since the patches for 2.4.20 seem to have been causing data corruption on unmount for XFS. Everything seems to be fine with the new ext3 + 2.4.20-ck3 combination that I use at home and in my workstation here at work. --> Jijo -- Federico Sevilla III : http://jijo.free.net.ph : When we speak of free Network Administrator : The Leather Collection, Inc. : software we refer to GnuPG Key ID : 0x93B746BE : freedom, not price. _ Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph To leave: send "unsubscribe" in the body to [EMAIL PROTECTED] Fully Searchable Archives With Friendly Web Interface at http://marc.free.net.ph To subscribe to the Linux Newbies' List: send "subscribe" in the body to [EMAIL PROTECTED]
