Hey Sasha,
Specifies the -l option in these respective scripts when they call
perfquery to clear counters/errors. This is the core of why I
implemented the -l option in perfquery. These two tools were failing in
a chunk of our switches.
Note that these scripts specify -R (reset only) for perfquery, so there
aren't any issues with multiple perfquery outputs that may need to be
parsed multiple times/differently.
Al
--
Albert Chu
[EMAIL PROTECTED]
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory
>From a8758c73b4c066aa8a68a08cfc7e4f1a382fe422 Mon Sep 17 00:00:00 2001
From: Albert Chu <[EMAIL PROTECTED]>
Date: Wed, 8 Oct 2008 17:28:16 -0700
Subject: [PATCH] specify --loop_ports when resetting counters/errors
Signed-off-by: Albert Chu <[EMAIL PROTECTED]>
---
infiniband-diags/scripts/ibclearcounters.in | 2 +-
infiniband-diags/scripts/ibclearerrors.in | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/infiniband-diags/scripts/ibclearcounters.in b/infiniband-diags/scripts/ibclearcounters.in
index 86a5528..d21fc6c 100644
--- a/infiniband-diags/scripts/ibclearcounters.in
+++ b/infiniband-diags/scripts/ibclearcounters.in
@@ -63,7 +63,7 @@ echo "$text" | awk '
function clear_counters(lid)
{
- if (system("'$IBPATH'/perfquery'"$ca_info"' '$gflags' -R -a " lid))
+ if (system("'$IBPATH'/perfquery'"$ca_info"' '$gflags' -R -a -l " lid))
nodeerr++
}
diff --git a/infiniband-diags/scripts/ibclearerrors.in b/infiniband-diags/scripts/ibclearerrors.in
index edf93f5..fefa0d3 100644
--- a/infiniband-diags/scripts/ibclearerrors.in
+++ b/infiniband-diags/scripts/ibclearerrors.in
@@ -67,7 +67,7 @@ echo "$text" | awk '
function clear_errors(lid, port)
{
- if (system("'$IBPATH'/perfquery'"$ca_info"' '$gflags' -R " lid " " port " 0x0fff"))
+ if (system("'$IBPATH'/perfquery'"$ca_info"' '$gflags' -R -l " lid " " port " 0x0fff"))
nodeerr++
}
--
1.5.4.5
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
