From: mails <ma...@da...> - 2013-09-30 17:30:12
|
Hi, I have a weird problem with smartd and temperature monitoring. Here is a simple test-configuration of mine: root@selektos:/tmp# cat /etc/smartd.conf.test /dev/sdb -W 0,20,20 -m XX...@gm... -M exec /usr/share/smartmontools/smartd-runner executing smartd -c /etc/smartd.conf.test does not send mail. if I execute /usr/share/smartmontools/smartd-runner by hand, I do get an empty email. So everything works fine on the mailsending part. What also works is the temperature logging in /var/log/smartd.log, but somehow smartd does not execute the mailsending script. Does anyone have a clue what might be the reason? I have this issue on a debian oldstable system. Smartd is version 5.40 Adrian |
From: Christian F. <Chr...@t-...> - 2013-09-30 19:34:36
|
mails wrote: > Hi, > > I have a weird problem with smartd and temperature monitoring. Here is a > simple test-configuration of mine: > > root@selektos:/tmp# cat /etc/smartd.conf.test > /dev/sdb -W 0,20,20 -m XX...@gm... -M exec > /usr/share/smartmontools/smartd-runner Please add '-M test' directive. A test mail is sent then regardless of any warning condition. See smartd.conf man page. > executing smartd -c /etc/smartd.conf.test does not send mail. > if I execute /usr/share/smartmontools/smartd-runner by hand, I do get an > empty email. So everything works fine on the mailsending part. > > What also works is the temperature logging in /var/log/smartd.log, but > somehow smartd does not execute the mailsending script. Please check these lines in smartd syslog output: Device: /dev/sdb, Temperature .. Celsius reached critical limit of 20 Celsius (Min/Max ../..) Sending warning via /usr/share/smartmontools/smartd-runner to ADDRESS ... Warning via /usr/share/smartmontools/smartd-runner to console: RESULT Which RESULT is printed? > > Does anyone have a clue what might be the reason? > I have this issue on a debian oldstable system. Smartd is version 5.40 This version of smartd is now 3 years old. I don't remember any similar error report. Please note that the smartd-runner script is Debian specific. Thanks, Christian |
From: mails <ma...@da...> - 2013-10-01 00:44:21
|
Hello Christian, thanks for the response. Am 30.09.2013 21:34, schrieb Christian Franke: > mails wrote: >> Hi, >> >> I have a weird problem with smartd and temperature monitoring. Here is a >> simple test-configuration of mine: >> >> root@selektos:/tmp# cat /etc/smartd.conf.test >> /dev/sdb -W 0,20,20 -m XX...@gm... -M exec >> /usr/share/smartmontools/smartd-runner > > Please add '-M test' directive. A test mail is sent then regardless of > any warning condition. See smartd.conf man page. I have tried that and the test directive works, but the test message gets sent only once after smartd's startup. so i cant use this as a work around. >> executing smartd -c /etc/smartd.conf.test does not send mail. >> if I execute /usr/share/smartmontools/smartd-runner by hand, I do get an >> empty email. So everything works fine on the mailsending part. >> >> What also works is the temperature logging in /var/log/smartd.log, but >> somehow smartd does not execute the mailsending script. > > Please check these lines in smartd syslog output: > > Device: /dev/sdb, Temperature .. Celsius reached critical limit of 20 > Celsius (Min/Max ../..) > Sending warning via /usr/share/smartmontools/smartd-runner to ADDRESS ... > Warning via /usr/share/smartmontools/smartd-runner to console: RESULT > > Which RESULT is printed? Maybe this is a point where to start looking. I only get the first line until (Min/Max ../..) >> >> Does anyone have a clue what might be the reason? >> I have this issue on a debian oldstable system. Smartd is version 5.40 > > This version of smartd is now 3 years old. I don't remember any similar > error report. > > Please note that the smartd-runner script is Debian specific. I understand that, but I am quite sure this is not where the problem lies. I changed the exec script to touch /tmp/FILE and no file was created. Greetings Adrian |
From: mails <ma...@da...> - 2013-10-01 10:30:37
|
Hello, I upgraded smartmontools to smartmontools_5.41 (squeeze-backports) too see if it solves my problem. I got one single notification after the installation. After that total silence again :( /var/log/smartd.log shows only one-liners like before. |
From: Christian F. <Chr...@t-...> - 2013-10-01 14:36:14
|
mails wrote: > Hello, > > I upgraded smartmontools to smartmontools_5.41 (squeeze-backports) too > see if it solves my problem. I got one single notification after the > installation. After that total silence again :( > /var/log/smartd.log shows only one-liners like before. > Please post these "one-liners". Does smart.log actually contain LOG_INFO and LOG_CRIT messages? (The message "Sending warning via ..." uses LOG_INFO level) |
From: mails <ma...@da...> - 2013-10-01 17:36:12
|
Am 01.10.2013 16:35, schrieb Christian Franke: > mails wrote: >> Hello, >> >> I upgraded smartmontools to smartmontools_5.41 (squeeze-backports) too >> see if it solves my problem. I got one single notification after the >> installation. After that total silence again :( >> /var/log/smartd.log shows only one-liners like before. >> > > Please post these "one-liners". > Does smart.log actually contain LOG_INFO and LOG_CRIT messages? > (The message "Sending warning via ..." uses LOG_INFO level) > The lines looks like this Oct 1 19:27:59 selektos smartd[1687]: Device: /dev/disk/by-id/wwn-0x5000c50029ea3503 [SAT], Temperature 35 Celsius reached critical limit of 26 Celsius (Min/Max 34/37) |
From: Christian F. <Chr...@t-...> - 2013-10-01 17:45:50
|
mails wrote: > Am 01.10.2013 16:35, schrieb Christian Franke: >> mails wrote: >>> Hello, >>> >>> I upgraded smartmontools to smartmontools_5.41 (squeeze-backports) too >>> see if it solves my problem. I got one single notification after the >>> installation. After that total silence again :( >>> /var/log/smartd.log shows only one-liners like before. >>> >> Please post these "one-liners". >> Does smart.log actually contain LOG_INFO and LOG_CRIT messages? >> (The message "Sending warning via ..." uses LOG_INFO level) >> > The lines looks like this > > Oct 1 19:27:59 selektos smartd[1687]: Device: > /dev/disk/by-id/wwn-0x5000c50029ea3503 [SAT], Temperature 35 Celsius > reached critical limit of 26 Celsius (Min/Max 34/37) > > LOG_INFO messages are missing in this log, this makes further diagnostics difficult. Please run "/usr/sbin/smartd -q onecheck" from root terminal and provide its standard output as an attachment. |
From: mails <ma...@da...> - 2013-10-01 17:50:30
Attachments:
onecheck.txt
|
Am 01.10.2013 19:45, schrieb Christian Franke: > mails wrote: >> Am 01.10.2013 16:35, schrieb Christian Franke: >>> mails wrote: >>>> Hello, >>>> >>>> I upgraded smartmontools to smartmontools_5.41 (squeeze-backports) too >>>> see if it solves my problem. I got one single notification after the >>>> installation. After that total silence again :( >>>> /var/log/smartd.log shows only one-liners like before. >>>> >>> Please post these "one-liners". >>> Does smart.log actually contain LOG_INFO and LOG_CRIT messages? >>> (The message "Sending warning via ..." uses LOG_INFO level) >>> >> The lines looks like this >> >> Oct 1 19:27:59 selektos smartd[1687]: Device: >> /dev/disk/by-id/wwn-0x5000c50029ea3503 [SAT], Temperature 35 Celsius >> reached critical limit of 26 Celsius (Min/Max 34/37) >> >> > > LOG_INFO messages are missing in this log, this makes further > diagnostics difficult. > > Please run "/usr/sbin/smartd -q onecheck" from root terminal and provide > its standard output as an attachment. > yes thats what i meant, with just one line. here is the standard output as an attachment. |
From: Christian F. <Chr...@t-...> - 2013-10-01 19:25:50
|
mails wrote: > > Am 01.10.2013 19:45, schrieb Christian Franke: >> Please run "/usr/sbin/smartd -q onecheck" from root terminal and provide >> its standard output as an attachment. >> > yes thats what i meant, with just one line. here is the standard output > as an attachment. This output suggests that there is not '-m' directive in the smartd.conf lines for the two devices. Did you possibly add entries after DEVICESCAN line? These are ignored, see man page. Please provide the smartd.conf file. |
From: mails <ma...@da...> - 2013-10-02 08:28:32
Attachments:
smartd.conf
|
Am 01.10.2013 21:25, schrieb Christian Franke: > mails wrote: >> >> Am 01.10.2013 19:45, schrieb Christian Franke: >>> Please run "/usr/sbin/smartd -q onecheck" from root terminal and provide >>> its standard output as an attachment. >>> >> yes thats what i meant, with just one line. here is the standard output >> as an attachment. > > This output suggests that there is not '-m' directive in the smartd.conf > lines for the two devices. > Did you possibly add entries after DEVICESCAN line? These are ignored, > see man page. > Please provide the smartd.conf file. thanks for looking into the matter. here is my smartd.conf file. There is no DEVICESCAN line. |
From: Christian F. <Chr...@t-...> - 2013-10-02 10:42:36
|
mails wrote: > Am 01.10.2013 21:25, schrieb Christian Franke: >> ... >> This output suggests that there is not '-m' directive in the smartd.conf >> lines for the two devices. >> Did you possibly add entries after DEVICESCAN line? These are ignored, >> see man page. >> Please provide the smartd.conf file. > thanks for looking into the matter. > here is my smartd.conf file. There is no DEVICESCAN line. > According to "smartd -q onecheck" output, state persistence (-s option) is used. Smartd older than r3278 use "-M once" as a default even in this case. This does not make sense because a warning message will never be resent. Workaround for older versions: Add "-M daily" to each config line. For testing, you could also edit the /var/lib/smartmontools/*.state files: remove "mail.*.* = " lines. Since r3429 (included in 5.42, ~2 years old), the warning condition is reset when temperature has dropped. A warning mail is resend immediately if the over temperature condition reappears. See http://sourceforge.net/apps/trac/smartmontools/ticket/167 Thanks, Christian |