summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authormickey <mickey@openbsd.org>2000-09-21 14:28:49 +0000
committermickey <mickey@openbsd.org>2000-09-21 14:28:49 +0000
commitbfb0773e7ac1fc13cb70455fe3f36ed9a95a7ee9 (patch)
tree7c361d27cc9bab3d33e042e9f27c9178061786bf
parentuse volatile's on all structures shared between cpu and hifn (diff)
downloadwireguard-openbsd-bfb0773e7ac1fc13cb70455fe3f36ed9a95a7ee9.tar.xz
wireguard-openbsd-bfb0773e7ac1fc13cb70455fe3f36ed9a95a7ee9.zip
add a firmware bug in raid0 or raid10 initialization; from http://people.freebsd.org/~msmith/RAID/
-rw-r--r--share/man/man4/twe.48
1 files changed, 7 insertions, 1 deletions
diff --git a/share/man/man4/twe.4 b/share/man/man4/twe.4
index c35ddb8442d..595448e8ab4 100644
--- a/share/man/man4/twe.4
+++ b/share/man/man4/twe.4
@@ -1,4 +1,4 @@
-.\" $OpenBSD: twe.4,v 1.4 2000/09/15 22:20:29 mickey Exp $
+.\" $OpenBSD: twe.4,v 1.5 2000/09/21 14:28:49 mickey Exp $
.\"
.\" Copyright (c) 2000 Michael Shalayeff. All rights reserved.
.\"
@@ -53,6 +53,12 @@ To avoid this situation, place
.Nm
cards in the lowest slot before other SCSI or RAID disk controllers.
.Pp
+The current 3ware firmware is known to fail if a redundant configuration
+is selected (RAID0 or RAID10) and the array is heavily used
+(eg. installation) before the controller has finished initialising the array.
+In this case, create the array, boot the installation media and
+leave the array idle untill all controller activity has ceased.
+.Pp
The card provides a mechanism to receive asynchronous notifications
about RAID set status change and so forth, unfortunately this
information is not public.