Closed Bug 92759 Opened 24 years ago Closed 23 years ago

SSL stress test hangs on Linux.

Categories

(NSS :: Tools, defect, P2)

Product:
NSS
Component:
Tools
Version:
3.2.1
Platform:
x86
Linux
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED INVALID

People

(Reporter: wtc, Assigned: kirk.erickson)

Details

I built NSS 3.3 on Red Hat Linux 7.1 and ran all.sh repeatedly. At one point the SSL stress test was hung and the 'ps' command's output was: wtc 13589 13580 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13600 13589 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13601 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13602 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13603 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13604 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13605 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13606 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13607 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13608 13600 0 19:11 pts/0 00:00:00 selfserv -D -p 8443 -d ../server wtc 13610 13580 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13613 13610 0 19:11 pts/0 00:00:00 [strsclnt <defunct>] wtc 13986 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13987 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13988 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13989 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13990 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13991 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss wtc 13992 1 0 19:11 pts/0 00:00:00 strsclnt -q -p 8443 -d . -w nss This bug may be related to bug 70643 in that the internal manager thread (pid 13613) for the strsclnt process (pid 13610) died (shown as 'defunct'). Apparently all the other threads in strsclnt had terminated and were adopted by the init process (pid 1) because their parent (the internal manager thread) was dead. The code was compiled on box.red.iplanet.com and the tests were run on nautica.red.iplanet.com.
Correction: all the other threads in strsclnt were still alive. They had not terminated, but they were adopted by the init process (pid 1).
Also note that pid 13580 is the process '/bin/sh ./all.sh'.
Similar hangs are seen every 100 runs on NT on the tinderbox
Priority: -- → P2
Target Milestone: --- → 3.4
Changed the QA contact to Bishakha.
QA Contact: sonja.mirtitsch → bishakhabanerjee
Set target milestone to NSS 3.5.
Target Milestone: 3.4 → 3.5
Set target milestone to NSS 3.6.
Target Milestone: 3.5 → 3.6
Let's look into this in 3.7. It's okay to resolve it INVALID or WONTFIX if we can't reproduce it on recent Linux releases (with the current kernel and glibc patches applied).
Target Milestone: 3.6 → 3.7
I've lost track of the version numbers. Here's what I have on box - my work machine: kernel-2.4.7-10 glibc-2.2.4-13 Is this sufficient?
I think those are the versions out of the box. You should go to http://www.redhat.com/apps/support/errata/ to download the latest kernel and glibc patches.
Do I need to force the install? or do I need to start gathering the dependent rpms? Downloaded from: http://mirror.hiwaay.net/redhat/redhat/linux/updates/7.2/en/os/i686/ Index of /redhat/redhat/linux/updates/7.2/en/os/i686 glibc-2.2.4-30.i686.rpm kernel-2.4.18-17.7.x.i686.rpm kernel-bigmem-2.4.18-17.7.x.i686.rpm kernel-debug-2.4.18-17.7.x.i686.rpm kernel-smp-2.4.18-17.7.x.i686.rpm [root@box 92759]# rpm -Uhv *.rpm error: failed dependencies: glibc-common = 2.2.4-30 is needed by glibc-2.2.4-30 modutils >= 2.4.18 is needed by kernel-2.4.18-17.7.x dev < 3.2-7 conflicts with kernel-2.4.18-17.7.x iptables < 1.2.5-3 conflicts with kernel-2.4.18-17.7.x bcm5820 < 1.81 conflicts with kernel-2.4.18-17.7.x modutils >= 2.4.18 is needed by kernel-bigmem-2.4.18-17.7.x dev < 3.2-7 conflicts with kernel-bigmem-2.4.18-17.7.x iptables < 1.2.5-3 conflicts with kernel-bigmem-2.4.18-17.7.x bcm5820 < 1.81 conflicts with kernel-bigmem-2.4.18-17.7.x modutils >= 2.4.18 is needed by kernel-debug-2.4.18-17.7.x dev < 3.2-7 conflicts with kernel-debug-2.4.18-17.7.x iptables < 1.2.5-3 conflicts with kernel-debug-2.4.18-17.7.x bcm5820 < 1.81 conflicts with kernel-debug-2.4.18-17.7.x modutils >= 2.4.18 is needed by kernel-smp-2.4.18-17.7.x dev < 3.2-7 conflicts with kernel-smp-2.4.18-17.7.x iptables < 1.2.5-3 conflicts with kernel-smp-2.4.18-17.7.x bcm5820 < 1.81 conflicts with kernel-smp-2.4.18-17.7.x
Do not force the install. You should install the dependent rpms, too.
Needed to download i386 rpms too. This combo installed on box: dev-3.3-4.i386.rpm glibc-2.2.4-30.i686.rpm glibc-common-2.2.4-30.i386.rpm glibc-devel-2.2.4-30.i386.rpm glibc-profile-2.2.4-30.i386.rpm hwcrypto-1.0-3.i386.rpm iptables-1.2.5-3.i386.rpm iptables-ipv6-1.2.5-3.i386.rpm kernel-2.4.18-17.7.x.i686.rpm kernel-bigmem-2.4.18-17.7.x.i686.rpm kernel-debug-2.4.18-17.7.x.i686.rpm kernel-smp-2.4.18-17.7.x.i686.rpm modutils-2.4.18-3.7x.i386.rpm modutils-devel-2.4.18-3.7x.i386.rpm openssl-0.9.6b-28.i386.rpm openssl-devel-0.9.6b-28.i386.rpm openssl-perl-0.9.6b-28.i386.rpm Added sonja.mirtitsch@sun.com to the list.
Sonja, let me know if you see the hangs on box go away. If the upgrade resolves this problem, we will mark this INVALID.
Status: NEW → ASSIGNED
I have not seen these hangs in a long time, and we are not supporting redhat 7.1 anymore. Please close.
Sonja reported that we're unable to reproduce.
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.