From: B. <fbo...@ca...> - 2006-05-16 14:52:56
|
Hello, I'm testing CCISS (RAID array) support patch of smartmontools through the smartmontools 5.36-6 Debian package on a HP ML350 G3 server, and I've noticed different problems (as described also in Debian Bug Tracking System, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D366802) : - the environment variables set by smartd when calling script through the -M option are incorrect : I can't call smartctl with them to retrieve disk information. - short and long tests scheduled on my 2 disks of the array are all done on the same disk (disk 1 ; disk 0 is never tested !) - since I've test smartd, an other array supervisor, cpqarrayd (which only monitor messages from the array), sometimes tells me, exactly one hour after the first short test (at 3h A.M.) : 03:25:17 cpqarrayd: CCISS controler /dev/cciss/c0d0 logical volume 0 change= d state to Logical drive is ready for recovery operation. 03:25:17 cpqarrayd: CCISS controler /dev/cciss/c0d0 logical volume 0 change= d state to Logical drive is is currently recovering. 03:25:17 cpqarrayd: CCISS controler /dev/cciss/c0d0 logical volume 0 change= d state to Logical drive is ok. I hope it can help testing this patch. With regards, Fr=E9d=E9ric Boiteux |
From: Bruce A. <ba...@gr...> - 2006-05-18 15:32:32
|
> - the environment variables set by smartd when calling script through > the -M option are incorrect : I can't call smartctl with them to > retrieve disk information. Can you please document this better? What have you done? Are the environment variables not set, or are they set to incorrect values? Cheers, Bruce |
From: Guido G. <ag...@si...> - 2006-05-19 05:35:43
|
Hi Bruce, On Thu, May 18, 2006 at 10:32:22AM -0500, Bruce Allen wrote: > >- the environment variables set by smartd when calling script through > >the -M option are incorrect : I can't call smartctl with them to > >retrieve disk information. > > Can you please document this better? What have you done? Are the > environment variables not set, or are they set to incorrect values? Most of the details are already here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=366802 For the envorinment vars the problem is that the envvars end up as: SMARTD_DEVICETYPE"="scsi" SMARTD_DEVICE"='/dev/cciss/c0d0 [cciss_disk_01]' I wonder if this is the same for the 3ware controllers? This is where the code is actually from. Didn't had a chance to look into the selftest problem but it would be nice if somebody could check if this only affects smartd or if starting selftests via smartctl is affected as well. Is so it'd be in the generic cciss code. Unfortunately I won't be near any cciss for the next 5 weeks. I'll have a look then if nobody else did until then. Please note that this affects no released version of smartmontools, only the latest version of the cciss patch that we circulated around here. Cheers, -- Guido |
From: B. <fbo...@ca...> - 2006-05-19 06:16:14
|
Hello Bruce and Guido, First, I wonder if I got right to post on smartmontools mailing list, as I didn't understood at first that CCISS wasn't an "official" addition to this project. Here are some more explanations about my problems : 1/ For the env. problem : Here is what I get in environment in the script called by smartd : SMARTD_TFIRSTEPOCH=3D1148018612 SMARTD_DEVICE=3D" SMARTD_DEVICESTRING=3D/dev/cciss/c0d0 [cciss_disk_00] SMARTD_MAILER=3D/etc/smartmontools/smartd-runner SMARTD_TFIRST=3DFri May 19 08:03:32 2006 CEST SMARTD_MESSAGE=3DTEST EMAIL from smartd for device: /dev/cciss/c0d0 [cciss_= disk_00] SMARTD_FAILTYPE=3DEmailTest SMARTD_ADDRESS=3D<my email> SMARTD_SUBJECT=3DSMART error (EmailTest) detected on host: <my host> SMARTD_DEVICETYPE=3Dscsi SMARTD_FULLMESSAGE=3DThis email was generated by the smartd daemon running = on: In my script (which run successfully on another disk), I do : /usr/sbin/smartctl -a -l selftest -l error -d "$SMARTD_DEVICETYPE" "$SMARTD= _DEVICE" To work, it should expand to : /usr/sbin/smartctl -a -l selftest -l error -d cciss,0 /dev/cciss/c0d0 but I get instead : /usr/sbin/smartctl -a -l selftest -l error -d scsi "/dev/cciss/c0d0 [cciss_= disk_00]" 2/ For smartd applying tests always on the same disk (always on disk 1, not 0, I want to add that using smartctl with same arguments, it works well, so it's a smartd problem only I think. 3/ at last, I didn't notice anymore any message from the cpqarrayd daemon (telling about recovering data from RAID 1 array) : I suspect that doing SMART tests reveal some bad blocks problems, automatically corrected by one disk, which causes the second to need to re-synchronize... with regards, Fr=E9d=E9ric Boiteux |
From: Bruce A. <ba...@gr...> - 2006-05-19 14:43:37
|
> First, I wonder if I got right to post on smartmontools mailing list, > as I didn't understood at first that CCISS wasn't an "official" > addition to this project. Guido is testing CCISS for inclusion in smartmontools. > Here is what I get in environment in the script called by smartd : > > SMARTD_TFIRSTEPOCH=1148018612 > SMARTD_DEVICE=" > SMARTD_DEVICESTRING=/dev/cciss/c0d0 [cciss_disk_00] > SMARTD_MAILER=/etc/smartmontools/smartd-runner > SMARTD_TFIRST=Fri May 19 08:03:32 2006 CEST > SMARTD_MESSAGE=TEST EMAIL from smartd for device: /dev/cciss/c0d0 [cciss_disk_00] > SMARTD_FAILTYPE=EmailTest > SMARTD_ADDRESS=<my email> > SMARTD_SUBJECT=SMART error (EmailTest) detected on host: <my host> > SMARTD_DEVICETYPE=scsi > SMARTD_FULLMESSAGE=This email was generated by the smartd daemon running on: It appears taht SMARTD_DEVICE and SMARTD_DEVICETYPE are not being set correctly by the CCISS patch. But I am not sure. I would need to see how they are set for the 3ware device type. That's the pattern to mimic. |
From: Volker K. <lis...@pa...> - 2006-05-20 02:58:08
|
On Fri 19 May 2006 18:15:59 NZST +1200, Fr=E9d=E9ric BOITEUX wrote: > First, I wonder if I got right to post on smartmontools mailing list, Everyone can send email to the list, but only subscribers get preferential treatment by having their emails posted automatically. The rest enters the moderation queue. I am the moderator, and when I started, I dealt with the queue once a day (notwithstanding time off =66rom the computer for a few days). It soon turned out that > 90% is spam, so I now look at it every 2-3 days. If you'd like your posts dealt with faster, or to save me some time, please consider subscribing: > Smartmontools-support mailing list > https://lists.sourceforge.net/lists/listinfo/smartmontools-support Because I don't like censorship, I approve everything which has to do with smartmontools. Spam stands out in the subject pretty well, but posting about smartmontools under a header of "how are you today" is taking risks. So far I am unaware of having made a mistake, but clicking the wrong button has happened to everyone before. Regards, Volker --=20 Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me. |
From: Bruce A. <ba...@gr...> - 2006-05-20 06:42:31
|
Volker, I wanted to publicly thank you for doing a nice job of administering the=20 list. I'm very happy with how you are doing this. Thank you! Cheers, =09Bruce On Sat, 20 May 2006, Volker Kuhlmann wrote: > On Fri 19 May 2006 18:15:59 NZST +1200, Fr=E9d=E9ric BOITEUX wrote: > >> First, I wonder if I got right to post on smartmontools mailing list, > > Everyone can send email to the list, but only subscribers get > preferential treatment by having their emails posted automatically. The > rest enters the moderation queue. I am the moderator, and when I > started, I dealt with the queue once a day (notwithstanding time off > from the computer for a few days). It soon turned out that > 90% is > spam, so I now look at it every 2-3 days. If you'd like your posts dealt > with faster, or to save me some time, please consider subscribing: > >> Smartmontools-support mailing list >> https://lists.sourceforge.net/lists/listinfo/smartmontools-support > > Because I don't like censorship, I approve everything which has to do > with smartmontools. Spam stands out in the subject pretty well, but > posting about smartmontools under a header of "how are you today" is > taking risks. So far I am unaware of having made a mistake, but clicking > the wrong button has happened to everyone before. > > Regards, > > Volker > > --=20 > Volker Kuhlmann=09=09=09is list0570 with the domain in header > http://volker.dnsalias.net/=09Please do not CC list postings to me. > > > ------------------------------------------------------- > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job ea= sier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronim= o > http://sel.as-us.falkag.net/sel?cmd______________________________________= _________ > Smartmontools-support mailing list > Sma...@li... > https://lists.sourceforge.net/lists/listinfo/smartmontools-support > |