| To: | Nathan Scott <nathans@xxxxxxx> |
|---|---|
| Subject: | Re: Data corruption with xfs+nfs+lvm |
| From: | Craig Tierney <ctierney@xxxxxxxx> |
| Date: | Fri, 13 Feb 2004 10:29:25 -0700 |
| Cc: | cattelan@xxxxxxx, linux-xfs@xxxxxxxxxxx |
| In-reply-to: | <1076360029.5321.53.camel@xxxxxxxxxxxxxxxxxx> |
| References: | <1075423747.3859.280.camel@xxxxxxxxxxxxxxxxxx> <20040130024343.GC1062@frodo> <1075482631.3866.9.camel@xxxxxxxxxxxxxxxxxx> <1076360029.5321.53.camel@xxxxxxxxxxxxxxxxxx> |
| Sender: | linux-xfs-bounce@xxxxxxxxxxx |
I wanted to update my investigation with the corruption I was seeing on my xfs filesystem exported over nfs. The problem is NOT xfs or nfs. It turns out that the tg3 driver in 2.4.20 and 2.4.21 doesn't hold up under load (at least this load). I have not tried a newer driver at this point, but switching to a Acenic based gigE card has solved the problem. I have not gone back to see what happens when I re-enable xfs_refcache_purge_some yet. We needed to get a project done. When that happens (in a few days) I can take that server back and enable that function and see if things still work. Craig |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: more experiments with realtime subvolumes., Eric Sandeen |
|---|---|
| Next by Date: | Re: more experiments with realtime subvolumes., Eric Sandeen |
| Previous by Thread: | Re: Data corruption with xfs+nfs+lvm, Craig Tierney |
| Next by Thread: | Corruption of in-memory data detected., Jeffrey W. Baker |
| Indexes: | [Date] [Thread] [Top] [All Lists] |