ZFS Monitoring Script
list Mike Rowell
Okay here it is, it's small but does the following. File system HEALTH. ONLINE = green DEGRADED = yellow FAULTED = red File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file. It doesn't check scrub status as yet as I'm not to sure what the scrub will report back, so once I've got some scrubbing underway I'll adjust the script to take into account scrubbing. This was something I threw together in 15minutes as we just realized that UFS would not suit the requirements of a project and therefore we had to move to ZFS. Enhancements I'm working on are to display all the storage devices in the pools and their health and also read / write and checksum errors, but they will have to wait for a bit. Mike Rowell Systems Team Leader * Email: user-63f3e97eb1de@xymon.invalid * Tel: XXX XXXX XXXX * Fax: XXXX XXX XXXX www.rightmove.co.uk <http://www.rightmove.co.uk/>; - the UK's number 1 property website This message (including any attachments) is confidential and may be legally privileged. The content and views expressed are those of the sender and not necessarily Rightmove plc. If you are not the intended recipient, you must not disclose, copy or use any part of it. Please delete all copies immediately and notify the sender. Rightmove plc is registered in England and Wales, No. 03997679. Registered Office: Rightmove plc, Grafton Court, Snowdon Drive, Milton Keynes, MK6 1AJ. This email has been scanned for all viruses by the MessageLabs service.
Attachments (1)
list Charles Goyard
Hi,
▸
Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.
I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards, -- Charles Goyard - user-a6cdca7046e2@xymon.invalid - (+33) 1 45 38 01 31 Orange Business Services - online multimedia // ingénierie
list Jason Altrincham Jones
Not sure if this has been covered previously but when you schedule a number of disables for the future then the only way to cancel the scheduled task seems to be to click cancel on each of them, it would be nice to have something similar to the "Currently disabled tests" listbox, if possible. If I am missing an easier way please tell me :-). Thanks, Jason.
list Rich Smrcina
I concur, unless there's something specific that ZFS provides that is critical to monitor. This way the back end client code will be similar to (or the same as) that for Linux.
▸
Charles Goyard wrote:Hi, Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards,
--
Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007
list Mike Rowell
Sorry made a typo when trying some new functionality, working version attached. Mike
▸
From: Mike Rowell [mailto:user-63f3e97eb1de@xymon.invalid] Sent: 12 April 2007 15:15 To: user-ae9b8668bcde@xymon.invalid Subject: [hobbit] ZFS Monitoring Script Okay here it is, it's small but does the following. File system HEALTH. ONLINE = green DEGRADED = yellow FAULTED = red File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file. It doesn't check scrub status as yet as I'm not to sure what the scrub will report back, so once I've got some scrubbing underway I'll adjust the script to take into account scrubbing. This was something I threw together in 15minutes as we just realized that UFS would not suit the requirements of a project and therefore we had to move to ZFS. Enhancements I'm working on are to display all the storage devices in the pools and their health and also read / write and checksum errors, but they will have to wait for a bit. Mike Rowell Systems Team Leader * Email: user-63f3e97eb1de@xymon.invalid * Tel: XXX XXXX XXXX * Fax: XXXX XXX XXXX www.rightmove.co.uk <http://www.rightmove.co.uk/>; - the UK's number 1 property website This message (including any attachments) is confidential and may be legally privileged. The content and views expressed are those of the sender and not necessarily Rightmove plc. If you are not the intended recipient, you must not disclose, copy or use any part of it. Please delete all copies immediately and notify the sender. Rightmove plc is registered in England and Wales, No. 03997679. Registered Office: Rightmove plc, Grafton Court, Snowdon Drive, Milton Keynes, MK6 1AJ. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service.
Attachments (1)
list Mike Rowell
The way ZFS works is to have storage pools, the script checks the status of the storage pools which is where the ONLINE, FAULTED, DEGRADED stuff comes in to it. The capacity is just for my own information as we currently run only one zfs filesystem per storage pool, as you state this isn't really needed as it does duplicate other things, however it gives you the ability to see how much storage is in a pool, you can have multiple ZFS file systems inside each pool so it does have it's uses. mike
▸
-----Original Message-----
From: Rich Smrcina [mailto:user-cf452ff334e0@xymon.invalid]
Sent: 12 April 2007 16:13
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] ZFS Monitoring Script
I concur, unless there's something specific that ZFS provides that is
critical to monitor. This way the back end client code will be similar
to (or the same as) that for Linux.
Charles Goyard wrote:Hi, Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards,
-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007 This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service.
list Greg Shea
Hi, Here's some output from one of my Sol10 boxes running ZFS. Don't quote me but I think the filesystems are carved up from the ZFS pools. -Grs- Gregory R Shea EMC Corporation [hobbit at solaris10 ]$ df -n -l / : ufs /devices : devfs /system/contract : ctfs /proc : proc /etc/mnttab : mntfs /etc/svc/volatile : tmpfs /system/object : objfs /dev/fd : fd /tmp : tmpfs /var/run : tmpfs /data : zfs /data/docp1 : zfs /data/docp1/dctm : zfs /data/docp1/dctm/data: zfs [hobbit at solaris10 ]$ /usr/sbin/zpool list -H node_apps 123M 6.83M 116M 5% ONLINE - ofcdocp1 7.94G 86K 7.94G 0% ONLINE - ofcdocp1_data 398G 86K 398G 0% ONLINE - ofcdocp1_doc 99.5G 88K 99.5G 0% ONLINE - [hobbit at solaris10 ]$ df -lk Filesystem kbytes used avail capacity Mounted on /dev/dsk/emcpower2a 12392428 4946137 7198443 41% / /devices 0 0 0 0% /devices ctfs 0 0 0 0% /system/contract proc 0 0 0 0% /proc mnttab 0 0 0 0% /etc/mnttab swap 6953160 1280 6951880 1% /etc/svc/volatile objfs 0 0 0 0% /system/object fd 0 0 0 0% /dev/fd swap 6951992 112 6951880 1% /tmp swap 6951928 48 6951880 1% /var/run node_apps 93184 6911 86204 8% /data ofcdocp1 8193024 25 8192941 1% /data/docp1 ofcdocp1_doc 102703104 26 102703019 1% /data/docp1/dctm ofcdocp1_data 410812416 24 410812333 1% /data/docp1/dctm/data
▸
-----Original Message-----
From: Rich Smrcina [mailto:user-cf452ff334e0@xymon.invalid]
Sent: Thursday, April 12, 2007 11:13 AM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] ZFS Monitoring Script
I concur, unless there's something specific that ZFS provides that is
critical to monitor. This way the back end client code will be similar
to (or the same as) that for Linux.
Charles Goyard wrote:Hi, Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards,
-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007
list T.J. Yang
As promised, here is hb client zfs wiki notes. http://en.wikibooks.org/wiki/System_Monitoring_with_Hobbit/HOWTO#Hobbit_Client_and_ZFS_monitoring It is not done yet, I haven't able to see the zfs column on hb server. T.J. Yang
▸
From: "Mike Rowell" <user-63f3e97eb1de@xymon.invalid> Reply-To: user-ae9b8668bcde@xymon.invalid To: <user-ae9b8668bcde@xymon.invalid> Subject: RE: [hobbit] ZFS Monitoring Script Date: Thu, 12 Apr 2007 16:13:43 +0100 Sorry made a typo when trying some new functionality, working version attached. Mike From: Mike Rowell [mailto:user-63f3e97eb1de@xymon.invalid] Sent: 12 April 2007 15:15 To: user-ae9b8668bcde@xymon.invalid Subject: [hobbit] ZFS Monitoring Script Okay here it is, it's small but does the following. File system HEALTH. ONLINE = green DEGRADED = yellow FAULTED = red File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file. It doesn't check scrub status as yet as I'm not to sure what the scrub will report back, so once I've got some scrubbing underway I'll adjust the script to take into account scrubbing. This was something I threw together in 15minutes as we just realized that UFS would not suit the requirements of a project and therefore we had to move to ZFS. Enhancements I'm working on are to display all the storage devices in the pools and their health and also read / write and checksum errors, but they will have to wait for a bit. Mike Rowell Systems Team Leader * Email: user-63f3e97eb1de@xymon.invalid * Tel: XXX XXXX XXXX * Fax: XXXX XXX XXXX www.rightmove.co.uk <http://www.rightmove.co.uk/>; - the UK's number 1 property website This message (including any attachments) is confidential and may be legally privileged. The content and views expressed are those of the sender and not necessarily Rightmove plc. If you are not the intended recipient, you must not disclose, copy or use any part of it. Please delete all copies immediately and notify the sender. Rightmove plc is registered in England and Wales, No. 03997679. Registered Office: Rightmove plc, Grafton Court, Snowdon Drive, Milton Keynes, MK6 1AJ. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service.
<< zfs_stat.sh >>
The average US Credit Score is 675. The cost to see yours: $0 by Experian. http://www.freecreditreport.com/pm/default.aspx?sc=660600&bcd=EMAILFOOTERAVERAGE
list Rich Smrcina
OK, then it's a software layer above the file systems, sort of like (but probably nothing like) LVM?
▸
Mike Rowell wrote:The way ZFS works is to have storage pools, the script checks the status of the storage pools which is where the ONLINE, FAULTED, DEGRADED stuff comes in to it. The capacity is just for my own information as we currently run only one zfs filesystem per storage pool, as you state this isn't really needed as it does duplicate other things, however it gives you the ability to see how much storage is in a pool, you can have multiple ZFS file systems inside each pool so it does have it's uses. mike -----Original Message----- From: Rich Smrcina [mailto:user-cf452ff334e0@xymon.invalid] Sent: 12 April 2007 16:13 To: user-ae9b8668bcde@xymon.invalid Subject: Re: [hobbit] ZFS Monitoring Script I concur, unless there's something specific that ZFS provides that is critical to monitor. This way the back end client code will be similar to (or the same as) that for Linux. Charles Goyard wrote:Hi, Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards,
-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007
list Mike Rowell
It's a bit of both... ZFS is a file system type that can be applied to a raw device to create a filesystem, however to get that you need to put the raw device into a zfs pool which is what the software part does. It's build directly into the kernel so operates in kernel space, but also has a lot of dynamic qualities such as compression that can be enabled and disabled at will, good if you've got a filesystem with a lot of logs in it etc. The good thing is that there is no such concept as inodes in zfs so you can have as many files as you want without worrying about hitting inodes maximums.
▸
Mike
-----Original Message-----
From: Rich Smrcina [mailto:user-cf452ff334e0@xymon.invalid]
Sent: 12 April 2007 16:51
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] ZFS Monitoring Script
OK, then it's a software layer above the file systems, sort of like (but
probably nothing like) LVM?
Mike Rowell wrote:The way ZFS works is to have storage pools, the script checks the status of the storage pools which is where the ONLINE, FAULTED, DEGRADED stuff comes in to it. The capacity is just for my own information as we currently run only one zfs filesystem per storage pool, as you state this isn't really needed as it does duplicate other things, however it gives you the ability to see how much storage is in a pool, you can have multiple ZFS file systems inside each pool so it does have it's uses. mike -----Original Message----- From: Rich Smrcina [mailto:user-cf452ff334e0@xymon.invalid] Sent: 12 April 2007 16:13 To: user-ae9b8668bcde@xymon.invalid Subject: Re: [hobbit] ZFS Monitoring Script I concur, unless there's something specific that ZFS provides that is critical to monitor. This way the back end client code will be similar to (or the same as) that for Linux. Charles Goyard wrote:Hi, Mike Rowell wrote :File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file.I don't know about ZFS, but are these the same kind of values "du" reports ? If so, maybe it's a duplicate, or maybe you should format the output the way "disk" does. Just a question/suggestion. Regards,
-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007 This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service.
list Rich Smrcina
Sounds pretty cool!
▸
Mike Rowell wrote:It's a bit of both... ZFS is a file system type that can be applied to a raw device to create a filesystem, however to get that you need to put the raw device into a zfs pool which is what the software part does. It's build directly into the kernel so operates in kernel space, but also has a lot of dynamic qualities such as compression that can be enabled and disabled at will, good if you've got a filesystem with a lot of logs in it etc. The good thing is that there is no such concept as inodes in zfs so you can have as many files as you want without worrying about hitting inodes maximums. Mike
-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid Catch the WAVV! http://www.wavv.org WAVV 2007 - Green Bay, WI - May 18-22, 2007
list Uwe Kirbach
Hello Mike + hobbits, i wanted to have a first line in the hobbit message, so that i can include this line in a sms alert. That's why i changed your script a little bit. It gives you "zfs - health: okay - capacity: okay" if all is okay Thanks Uwe Kirbach EnBW Energie Baden-Württemberg Systeme Infrastruktur Support GmbH Informationsverarbeitung Operations Infrastrukturbetrieb Unix Sitz der Gesellschaft: Karlsruhe Handelsregister: Amtsgericht Mannheim - HRB 108550 Vorsitzender des Aufsichtsrats: Dr. Bernhard Beck Geschäftsführer: Jochen Adenau, Dr. Peter Krampf
▸
-----Original Message----- From: Mike Rowell [mailto:user-63f3e97eb1de@xymon.invalid] Sent: Thursday, April 12, 2007 5:14 PM To: user-ae9b8668bcde@xymon.invalid Subject: RE: [hobbit] ZFS Monitoring Script Sorry made a typo when trying some new functionality, working version attached. Mike From: Mike Rowell [mailto:user-63f3e97eb1de@xymon.invalid] Sent: 12 April 2007 15:15 To: user-ae9b8668bcde@xymon.invalid Subject: [hobbit] ZFS Monitoring Script Okay here it is, it's small but does the following. File system HEALTH. ONLINE = green DEGRADED = yellow FAULTED = red File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file. It doesn't check scrub status as yet as I'm not to sure what the scrub will report back, so once I've got some scrubbing underway I'll adjust the script to take into account scrubbing. This was something I threw together in 15minutes as we just realized that UFS would not suit the requirements of a project and therefore we had to move to ZFS. Enhancements I'm working on are to display all the storage devices in the pools and their health and also read / write and checksum errors, but they will have to wait for a bit. Mike Rowell Systems Team Leader * Email: user-63f3e97eb1de@xymon.invalid * Tel: XXX XXXX XXXX * Fax: XXXX XXX XXXX www.rightmove.co.uk <http://www.rightmove.co.uk/>; - the UK's number 1 property website This message (including any attachments) is confidential and may be legally privileged. The content and views expressed are those of the sender and not necessarily Rightmove plc. If you are not the intended recipient, you must not disclose, copy or use any part of it. Please delete all copies immediately and notify the sender. Rightmove plc is registered in England and Wales, No. 03997679. Registered Office: Rightmove plc, Grafton Court, Snowdon Drive, Milton Keynes, MK6 1AJ. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service. This email has been scanned for all viruses by the MessageLabs service.
Attachments (1)
list Galen Johnson
▸
T.J. Yang wrote:
As promised, here is hb client zfs wiki notes. http://en.wikibooks.org/wiki/System_Monitoring_with_Hobbit/HOWTO#Hobbit_Client_and_ZFS_monitoring It is not done yet, I haven't able to see the zfs column on hb server. T.J. YangFrom: "Mike Rowell" <user-63f3e97eb1de@xymon.invalid> Reply-To: user-ae9b8668bcde@xymon.invalid To: <user-ae9b8668bcde@xymon.invalid> Subject: RE: [hobbit] ZFS Monitoring Script Date: Thu, 12 Apr 2007 16:13:43 +0100 Sorry made a typo when trying some new functionality, working version attached. Mike From: Mike Rowell [mailto:user-63f3e97eb1de@xymon.invalid] Sent: 12 April 2007 15:15 To: user-ae9b8668bcde@xymon.invalid Subject: [hobbit] ZFS Monitoring Script Okay here it is, it's small but does the following. File system HEALTH. ONLINE = green DEGRADED = yellow FAULTED = red File system sizes are set to alert yellow at 80% capacity and red at 90%, the variables to alter this are at the top of the file. It doesn't check scrub status as yet as I'm not to sure what the scrub will report back, so once I've got some scrubbing underway I'll adjust the script to take into account scrubbing. This was something I threw together in 15minutes as we just realized that UFS would not suit the requirements of a project and therefore we had to move to ZFS. Enhancements I'm working on are to display all the storage devices in the pools and their health and also read / write and checksum errors, but they will have to wait for a bit.
<< zfs_stat.sh >>
Let's not forget about The Shire folks... http://www.trantor.org/theshire/doku.php/faq has a template you can send to the developers list on the Shire... =G=