[OpenAFS-devel] Problems with 1.4.8pre2 on Fedora 8

[Harald Barth]

----Next_Part(Wed_Oct_15_11_50_11_2008_966)--
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit


Unfortunately I have "still no cigar". But here is some more info that
might help you:

* I patched afs_BlackListOnce() and its invocation a little bit more
  because I think as is it allways returns 1. Se attached patch.

* If I turn on fstrace, I can not repeat the timeout any more,
  probably the client is to slow to trigger the bug then.

Other guess: How dangerous is the new goto out in afs_Analyze() as it
bypasses a lot of code?

Back to logfile debugging after lunch,
Harald.


----Next_Part(Wed_Oct_15_11_50_11_2008_966)--
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline; filename="openafs-1.4.8pre2.serverdown.patch"

--- openafs-1.4.8pre2/src/afs/afs_analyze.c.orig	2008-08-22 20:53:16.000000000 +0200
+++ openafs-1.4.8pre2/src/afs/afs_analyze.c	2008-10-15 10:19:08.000000000 +0200
@@ -455,7 +455,7 @@
  *              blacklist. 
  *
  * Returns:
- *	Non-zero value if further servers are available to try,
+ *	The number of further servers available to try,
  *	zero otherwise.
  *
  * Environment:
@@ -485,7 +485,7 @@
 		    areq->skipserver[i] = 1;
 		}
 		if (tvp->serverHost[i] &&
-		    !(tvp->serverHost[i]->addr->sa_flags & 
+		    (tvp->serverHost[i]->addr->sa_flags & 
 		      SRVR_ISDOWN)) {
 		    areq->skipserver[i] = 1;
 		}
@@ -494,9 +494,8 @@
 	}
     }
     for (i = 0; i < MAXHOSTS; i++) {
-	if (areq->skipserver[i] == 0) {
-	    serversleft = 1;
-	    break;
+	if (tvp->serverHost[i] && areq->skipserver[i] == 0) {
+	    serversleft++;
 	}
     }
     return serversleft;
@@ -545,7 +544,6 @@
     struct server *tsp;
     struct volume *tvp;
     afs_int32 shouldRetry = 0;
-    afs_int32 serversleft = 1;
     struct afs_stats_RPCErrors *aerrP;
     afs_int32 markeddown;
 
@@ -672,9 +670,8 @@
 #endif /* AFS_64BIT_CLIENT */
     if ((acode < 0) && (acode != VRESTARTING)) {
 	if (acode == RX_CALL_TIMEOUT) {
-	    serversleft = afs_BlackListOnce(areq, afid, tsp);
 	    areq->idleError++;
-	    if (serversleft) {
+	    if (afs_BlackListOnce(areq, afid, tsp)) {
 		shouldRetry = 1;
 	    } else {
 		shouldRetry = 0;
@@ -746,10 +743,8 @@
 		    ("afs: Tokens for user of AFS id %d for cell %s have expired\n",
 		     tu->vid, aconn->srvr->server->cell->cellName);
 	    } else {
-		serversleft = afs_BlackListOnce(areq, afid, tsp);
 		areq->tokenError++;
-
-		if (serversleft) {
+		if (afs_BlackListOnce(areq, afid, tsp)) {
 		    afs_warnuser
 			("afs: Tokens for user of AFS id %d for cell %s: rxkad error=%d\n",
 			 tu->vid, aconn->srvr->server->cell->cellName, acode);

----Next_Part(Wed_Oct_15_11_50_11_2008_966)----