Bug#349287: am-utils: amd segfaulting on amd64 (x86_64) again
Package: am-utils Version: 6.2+rc20110530-3 Followup-For: Bug #349287 Dear Maintainer, *** Please consider answering these questions, where appropriate *** * am-utils worked fine for years until my last update * I noticed the problem since the last update on 013-07-21 (Brought new Kernel 3.9, ...) * When I do /etc/init.d/am-utils start I see in /var/log/messages: Jun 21 12:23:14 mother kernel: [12795.704797] amd[9167]: segfault at 0 ip 7f39ee202c13 sp 73b4ec50 error 4 in libc-2.17.so[7f39ee1c8000+1a4000] Jun 21 12:23:14 mother kernel: [12795.705343] amd[9168]: segfault at 0 ip 7f39ee202c13 sp 73b4ec50 error 4 in libc-2.17.so[7f39ee1c8000+1a4000] *** End of the template - remove these lines *** -- System Information: Debian Release: jessie/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 3.9-1-amd64 (SMP w/8 CPU cores) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages am-utils depends on: ii debconf1.5.50 ii debianutils4.3.4 ii libamu46.2+rc20110530-3 ii libc6 2.17-3 ii libgdbm3 1.8.3-12 ii libhesiod0 3.2.1-1 ii libldap-2.4-2 2.4.31-1+nmu2 ii libwrap0 7.6.q-24 ii rpcbind [portmap] 0.2.0-8 ii ucf3.0027 am-utils recommends no packages. Versions of packages am-utils suggests: ii am-utils-doc 6.2+rc20110530-3 pn nis none -- Configuration Files: /etc/am-utils/amd.conf changed: [global] ### Global parameters # Override the achitecture ? #arch = i386 # Where to mount the filesystems auto_dir = /amd # Override the default 300 seconds (5 minutes) cache timeout #cache_duration = 300 # Set a cluster string ? #cluster = lab # Debug ? #debug_options = all # Override the default 120 seconds (2 minutes) umount interval #dismount_interval = 120 # Override the full OS string #full_os = linux-2.2.18 # Do we want to reference hosts via their FQDN ? #fully_qualified_hosts = no # What's the hesiod base name ? #hesiod_base = automount # Kernel architecture override #karch = i386 # LDAP parameters #ldap_base = #ldap_cache_maxmem = 131072 #ldap_cache_seconds = 0 #ldap_hostPorts = # Override the DNS domainname (defaults to the part after the first # dot of the machine fully qualified domain name (FQDN) #local_domain = yourdomain.org # Where to log to... log_file = syslog # What to log ? log_options = all,noinfo,nostats,nomap # What IP protocol to use when doing NFS ? (tcp or udp) # If specified here, *all* mounts will use udp, regardless of the map contents #nfs_proto = udp # How many retransmissions will the kernel attempt when communicating # with amd ? #nfs_restransmit_counter = 11 # Timeout in tenths of second between retransmissions when the kernel # communicates with amd ? #nfs_retry_interval = 8 # NFS version to use ? (2 or 3) # If specified here, *all* mounts will use it, regardless of the map contents #nfs_vers = 2 # NIS domain name to retrieve map from (defaults to the locally bound # domain) #nis_domain = your-nis-domain # Do we want to normalize host names when expanding ${rhost} #normalize_hostnames = no # Override the os string ? #os = linux # Override the os version ? #osver = 2.2.18 # Where to print the PID if print_pid is set #pid_file = /dev/stdout # Try to lock amd into memory with plock() ? #plock = yes # What portmapper program number do we register for communications between # amd and the various utilities ? #portmap_program = 300019 # Print our pid on startup ? (not needed, we use amq for that) #print_pid = no # Print amd version on startup ? #print_version = no # At startup, do we restart existing mounts if we determine they could # have been automounted ? restart_mounts = yes # Do we want to be able to use selectors on the /defaults map entry ? selectors_in_defaults = yes # Do we return the number of automount entries on statfs()/df ? #show_statfs_entries = no # Do we attempt to unmount all automounted file systems on exit ? unmount_on_exit = yes # Override vendor ? vendor = Debian ### Default parameters, can be overriden on a map-by-map basis # Do we show unmounted automount points on readdir() browsable_dirs = yes # Default map options #map_options = # Map type ? Which kinds of map to look for (can be: file, hesiod, ldap, # ndbm, nis, nisplus, passwd, union), defaults to all of these. #map_type = # Default Mount type ? #mount_type = nfs # Path to search maps into (for files or ndbm maps) ? #search_path = [/mpnet] map_name = /etc/am-utils/mpnet -- debconf information: am-utils/import-amd-failed: am-utils/nis-master-map: amd.master am-utils/clustername: am-utils/nis-key: default am-utils/rpc-localhost:
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
Did you try to run the amd process under valgrind and wait until it crashed? valgrind is very good at reporting segfaults. It would be even better if you build the binary with debug symlinks, to get function names instead of addresses from the backtrace. Changing the init.d script like this would run the amd program under valgrind and lot any issues into /tmp/amd-log*. diff -ur ../am-utils-6.1.5/debian/am-utils.init ../am-utils-6.1.5-pere/debian/am-utils.init --- ../am-utils-6.1.5/debian/am-utils.init 2009-09-14 10:10:50.0 +0200 +++ ../am-utils-6.1.5-pere/debian/am-utils.init 2009-09-14 10:27:46.0 +0200 @@ -113,7 +113,7 @@ get_amd_args echo -n Starting automounter: amd -/usr/sbin/amd -F /etc/am-utils/amd.conf $dnsdomain $AMDARGS +valgrind --log-file=/tmp/amd-log /usr/sbin/amd -F /etc/am-utils/amd.conf $dnsdomain $AMDARGS echo . } Happy hacking, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
On 24 Jan 2006, at 1:04 am, John Gruenenfelder wrote: Oh, I should have made that a little more clear. The machine in question is not actually using the amd.home map. The other two machines in the cluster use it. I just included it for completeness. On this particular machine, /home is local so there is no need for a map. But, to insure that some file is accessible on all machines through the same path, we often use things like /net/bach/d1/foo even when on machine bach. That way the path doesn't change if accessed from another machine. Right - I'm doing something similar here, although a slightly different way. There have been some reports on the am-utils mailing list about problems with the link mount type on debian; you were the first but you're not the only one. There may be a bug, but no-one's managed to identify it yet. I have just uploaded the latest stable version (6.1.5) to debian unstable. You'll have to wait for the AMD64 autobuilder to do its thing, so it may take a day or two to appear. I'd appreciate it if you could have a look at it and see if the problem is still there. If it is, then we'll have to get some more debugging information to the am-utils developers. Tim -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
Thanks for the files. I can't currently see how you're telling amd to use your amd.home map; it's not mentioned in the defaults/am-utils file, and it's commented out in the amd.conf file - did you modify the /etc/init.d/am-utils file directly to do it, and if so, can you tell me what the exact command line is that am-utils is getting launched with? Tim -- Dr Tim Cutts Informatics Systems Group, Wellcome Trust Sanger Institute GPG: 1024D/E3134233 FE3D 6C73 BBD6 726A A3F5 860B 3CDD 3F56 E313 4233 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
On Mon, Jan 23, 2006 at 09:15:07AM +, Tim Cutts wrote: Thanks for the files. I can't currently see how you're telling amd to use your amd.home map; it's not mentioned in the defaults/am-utils file, and it's commented out in the amd.conf file - did you modify the /etc/init.d/am-utils file directly to do it, and if so, can you tell me what the exact command line is that am-utils is getting launched with? Tim Oh, I should have made that a little more clear. The machine in question is not actually using the amd.home map. The other two machines in the cluster use it. I just included it for completeness. On this particular machine, /home is local so there is no need for a map. But, to insure that some file is accessible on all machines through the same path, we often use things like /net/bach/d1/foo even when on machine bach. That way the path doesn't change if accessed from another machine. The other two machines which use amd.home receive it via NIS. -- --John GruenenfelderResearch Assistant, UMass Amherst student Systems Manager, MKS Imaging Technology, LLC. Try Weasel Reader for PalmOS -- http://gutenpalm.sf.net This is the most fun I've had without being drenched in the blood of my enemies! --Sam of Sam Max -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
On 22 Jan 2006, at 1:14 am, John Gruenenfelder wrote: Package: am-utils Version: 6.1.3-2 Severity: important amd has been exhibiting some nasty problems on an amd64 machine running the x86_64 port of Debian/testing. Specifically, it periodically segfaults and must be manually restarted. The usage of amd on this machine is important, but the machine is currently under a very light load. This is interesting - I'm also running am-utils on x86_64, but not this particular version (my production machines are still running the version in sarge). I see many messages in the log like this: Jan 21 16:07:17 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:07:18 bach kernel: nfs: server [EMAIL PROTECTED]:/net OK Jan 21 16:07:27 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:07:27 bach kernel: nfs: server [EMAIL PROTECTED]:/net OK Jan 21 16:26:58 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:26:58 bach kernel: amd[2840]: segfault at rip 2b1a3890 rsp 7fe93c08 error 4 The not responding/OK messages occur once or twice an hour. Currently, the other machines in our small cluster are not being used, so all of the NFS traffic is local to this machine. Can you actually reproduce this on demand, or does it just occasionally fall over? Could add your automount map files to the bug report, so that I (and the am-utils developers) can see exactly what the configuration is? I've seen some reports similar to this on the am-utils mailing list, so I'll forward your bug report to them and see what they think. But in the mean time, the full details of your configuration would be useful, i.e. your amd.conf file, your /etc/default/am-utils file and any map files referred to by the amd.conf file. Thanks, Tim -- Dr Tim Cutts Informatics Systems Group, Wellcome Trust Sanger Institute GPG: 1024D/E3134233 FE3D 6C73 BBD6 726A A3F5 860B 3CDD 3F56 E313 4233 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
On Sun, Jan 22, 2006 at 12:28:16PM +, Tim Cutts wrote: Jan 21 16:26:58 bach kernel: amd[2840]: segfault at rip 2b1a3890 rsp 7fe93c08 error 4 The not responding/OK messages occur once or twice an hour. Currently, the other machines in our small cluster are not being used, so all of the NFS traffic is local to this machine. Can you actually reproduce this on demand, or does it just occasionally fall over? Could add your automount map files to the bug report, so that I (and the am-utils developers) can see exactly what the configuration is? Unfortunately, I can't reproduce this on demand. At least, I haven't discovered if there is some specific thing that a user can do to trigger it. It seems to happen more or less at random. As I mentioned above, the not responding/OK messages occur once or twice and hour. The segfaulting seems to occur about once every 1-2 days. Before this machine's hardware was upgraded (it used to be x86) and Debian reinstalled, the not responding/OK messages occured, but with less regularity. amd did not segfault, however, and NFS seemed okay, so I just ignored the messages. I've seen some reports similar to this on the am-utils mailing list, so I'll forward your bug report to them and see what they think. But in the mean time, the full details of your configuration would be useful, i.e. your amd.conf file, your /etc/default/am-utils file and any map files referred to by the amd.conf file. Thanks, Tim Can do. I've attached my amd.conf and /etc/default/am-utils files as well as the amd.home map. -- --John GruenenfelderResearch Assistant, UMass Amherst student Systems Manager, MKS Imaging Technology, LLC. Try Weasel Reader for PalmOS -- http://gutenpalm.sf.net This is the most fun I've had without being drenched in the blood of my enemies! --Sam of Sam Max # Sample amd.conf for Debian GNU/Linux # $Id: amd.conf,v 1.5 2002/03/20 19:16:29 phil Exp $ # For the commented options, the default is shown [global] ### Global parameters # Override the achitecture ? #arch = i386 # Where to mount the filesystems auto_dir = /amd # Override the default 300 seconds (5 minutes) cache timeout #cache_duration = 300 # Set a cluster string ? #cluster = lab # Debug ? #debug_options = all # Override the default 120 seconds (2 minutes) umount interval #dismount_interval = 120 # Override the full OS string #full_os = linux-2.2.18 # Do we want to reference hosts via their FQDN ? #fully_qualified_hosts = no # What's the hesiod base name ? #hesiod_base = automount # Kernel architecture override #karch = i386 # LDAP parameters #ldap_base = #ldap_cache_maxmem = 131072 #ldap_cache_seconds = 0 #ldap_hostPorts = # Override the DNS domainname (defaults to the part after the first # dot of the machine fully qualified domain name (FQDN) #local_domain = yourdomain.org # Where to log to... log_file = syslog # What to log ? log_options = all,noinfo,nostats,nomap # What IP protocol to use when doing NFS ? (tcp or udp) # If specified here, *all* mounts will use udp, regardless of the map contents # nfs_proto = udp # How many retransmissions will the kernel attempt when communicating # with amd ? #nfs_restransmit_counter = 11 # Timeout in tenths of second between retransmissions when the kernel # communicates with amd ? #nfs_retry_interval = 8 # NFS version to use ? (2 or 3) # If specified here, *all* mounts will use it, regardless of the map contents nfs_vers = 3 # NIS domain name to retrieve map from (defaults to the locally bound # domain) #nis_domain = your-nis-domain # Do we want to normalize host names when expanding ${rhost} #normalize_hostnames = no # Override the os string ? #os = linux # Override the os version ? #osver = 2.2.18 # Where to print the PID if print_pid is set #pid_file = /dev/stdout # Try to lock amd into memory with plock() ? #plock = yes # What portmapper program number do we register for communications between # amd and the various utilities ? #portmap_program = 300019 # Print our pid on startup ? (not needed, we use amq for that) #print_pid = no # Print amd version on startup ? #print_version = no # At startup, do we restart existing mounts if we determine they could # have been automounted ? restart_mounts = yes # Do we want to be able to use selectors on the /defaults map entry ? selectors_in_defaults = yes # Do we return the number of automount entries on statfs()/df ? #show_statfs_entries = no # Do we attempt to unmount all automounted file systems on exit ? unmount_on_exit = yes # Override vendor ? vendor = Debian ### Default parameters, can be overriden on a map-by-map basis # Do we show unmounted automount points on readdir() #browsable_dirs = yes # Default map options
Bug#349287: am-utils: amd segfaulting on amd64 (x86_64)
Package: am-utils Version: 6.1.3-2 Severity: important amd has been exhibiting some nasty problems on an amd64 machine running the x86_64 port of Debian/testing. Specifically, it periodically segfaults and must be manually restarted. The usage of amd on this machine is important, but the machine is currently under a very light load. I see many messages in the log like this: Jan 21 16:07:17 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:07:18 bach kernel: nfs: server [EMAIL PROTECTED]:/net OK Jan 21 16:07:27 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:07:27 bach kernel: nfs: server [EMAIL PROTECTED]:/net OK Jan 21 16:26:58 bach kernel: nfs: server [EMAIL PROTECTED]:/net not responding, still trying Jan 21 16:26:58 bach kernel: amd[2840]: segfault at rip 2b1a3890 rsp 7fe93c08 error 4 The not responding/OK messages occur once or twice an hour. Currently, the other machines in our small cluster are not being used, so all of the NFS traffic is local to this machine. This configuration was previously used for quite some time on our old hardware which was x86. On the new amd64 machine this has been run under kernel 2.6.12 and 2.6.15 with similar results. -- System Information: Debian Release: testing/unstable APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.15 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages am-utils depends on: ii debconf [debconf-2.0] 1.4.67 Debian configuration management sy ii libamu4 6.1.3-2Support library for amd the 4.4BSD ii libc6 2.3.5-8.1 GNU C Library: Shared libraries an ii libhesiod03.0.2-16 Libraries for hesiod, a service na ii libldap2 2.1.30-12 OpenLDAP libraries ii libwrap0 7.6.dbs-8 Wietse Venema's TCP wrappers libra ii perl 5.8.7-10 Larry Wall's Practical Extraction ii portmap 5-16 The RPC portmapper am-utils recommends no packages. -- debconf information: am-utils/import-amd-failed: am-utils/nis-master-map: amd.master am-utils/config-reimport: import am-utils/nis-key: default * am-utils/done: yes am-utils/map-others: am-utils/nis-master-map-key-style: onekey am-utils/config-md5sum: bf8b03bb9b47986ec8d2dbd8690a85db * am-utils/map-net: true am-utils/config-reimport-failed: am-utils/reconfigure: * am-utils/use-nis: false am-utils/nis-custom: echo /amd-is-misconfigured /usr/share/am-utils/amd.net am-utils/import-amd-conf-done: false am-utils/import-amd-conf: false * am-utils/map-home: false am-utils/log-to-file: -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]