|latest/stable||dc935e5||18||28 Apr 2022|
juju deploy hw-health
This charm installs various hardware system monitoring tools and configures Nagios NRPE checks. It will only work for bare-metal installations on specific hardware.
Currently supported hardware is:
- Dell: PERC (perccli)
- Any other controller supported by the
- Supermicro: LSI SAS3008 RAID card with sas3ircu (Broadcoam's SAS3IRCU_P16)
- Huawei: LSI SAS2308 RAID card with sas2ircu (Huawei FusionServer Tools InfoCollect)
- SSD cards: Intel's PCIe Data Center SSD, Samsung's NVMe controllers for SM961/PM961 and 172Xa/172Xb.
- Linux software RAID (mdadm)
- IPMI as implemented by freeipmi (enable_ipmi config option is enabled by default)
- ipmiseld from the freeipmi suite (enable_ipmiseld is enabled by default) for logging system event log entries to syslog
In the backlog, hp-health logic still needs to be backported to support Hewlett-Packard gen8 and older equipment (HP Controllers with hpacucli)
Furthermore, other hardware in the roadmap is:
- Huawei's ES3000 V2 PCIe SSD Card with hio_info (Huawei ES3000 V2 Driver)
- S.M.A.R.T. Monitoring tool (smartctl)
juju deploy ubuntu juju deploy hw-health juju deploy nrpe juju add-relation ubuntu nrpe juju add-relation ubuntu hw-health juju add-relation hw-health nrpe
The Charmstore version already ships a resource. However, this resource is empty to avoid violating software redistribution license issues. To be useful, a new resource must be attached that includes your hardware manufacturer's RAID tools:
- Option 1:
juju deploy hw-health --resource tools=/tmp/zipfile.zip
- Option 2:
juju attach-resource hw-health tools=/tmp/zipfile.zip
In both cases format of zipfile.zip must be one of the following:
zip /tmp/zipfile.zip megacli sas2ircu sas3ircu perccli zip /tmp/zipfile.zip megacli etc.
SEL entries can be filtered by date, in order to allow to maintain monitoring SEL content without the need to clear it.
To filter out all current SEL entries you must use the
ack-sel action. This
will leave out from the IPMI check all SEL entries older than today.
ack-sel action optionally takes a
date parameter. SEL entries older
date will be ignored in the check.
show-sel action also obeys the date filter.
juju run-action hw-health/8 ack-sel --wait # or juju run-action hw-health/8 ack-sel date=2019-08-24 --wait # view juju run-action hw-health/8 show-sel --wait
To clear the filter (ie., consider all SEL entries present), you must use the
Under the hood, the filtering is done by appending a
parameter to the
check_ipmi_sensor NRPE plugin. The charm will do the right
thing if a
--seloptions parameter is already present via the
ipmi_check_options config. But the SEL filtering set by the
action will take precedence over a date filter set manually by the
ipmi_check_options config. This is:
$ date Wed Mar 10 15:36:36 UTC 2021 $ juju config hw-health/0 ipmi_check_options='--seloptions --date-range=07/02/2019-now' $ juju run-action hw-health/0 ack-sel
... will cause SEL entries older than the 10th of March 2021 to be ignored.
Known Limitations and Issues
Charm only install method is via Juju resources. There are plans to support snaps but snapstore only supports strictly confined snaps. Hardware monitoring tools need special permissions that are under development.
"tools" resource needs to be attached in ZIP format, and hardware monitoring tool(s) need to be on the first level of the archive tree.
Building the tools.zip resource
In order to build the tools.zip resource it is necessary to source the binaries from the respective vendor support pages.
For example: megacli/ sas3ircu/ sas2ircu/ [perccli] (https://dl.dell.com/FOLDER07576010M/1/PERCCLI_7.1623.00_A11_Linux.tar.gz)
You will then have to extract, rename, and compress the binaries to obtain the following structure:
$ zipinfo tools.zip Archive: tools.zip Zip file size: 3925441 bytes, number of entries: 4 -rwxr-xr-x 3.0 unx 2720320 bx defN 19-Jan-16 21:31 megacli -rwxr-xr-x 3.0 unx 7791416 bx defN 21-Oct-29 12:23 perccli -rwxrwxr-x 3.0 unx 559164 bx defN 19-Jan-16 21:31 sas2ircu -rwxrwxr-x 3.0 unx 562560 bx defN 19-Jan-16 21:31 sas3ircu 4 files, 11633460 bytes uncompressed, 3924843 bytes compressed: 66.3%
Two more zip resources may be needed for functional tests to succeed:
- tools-checksum.zip replaces the megacli tool by an empty file.
- tools-missing.zip removes the megacli tool from the resource
$ zipinfo tools-checksum.zip Archive: tools-checksum.zip Zip file size: 3239407 bytes, number of entries: 4 -rw-rw-r-- 3.0 unx 0 bx stor 21-Oct-29 16:04 megacli -rwxr-xr-x 3.0 unx 7791416 bx defN 21-Oct-29 12:23 perccli -rwxrwxr-x 3.0 unx 559164 bx defN 19-Jan-16 21:31 sas2ircu -rwxrwxr-x 3.0 unx 562560 bx defN 19-Jan-16 21:31 sas3ircu 4 files, 8913140 bytes uncompressed, 3238809 bytes compressed: 63.7% $ zipinfo tools-missing.zip Archive: tools-missing.zip Zip file size: 3239265 bytes, number of entries: 3 -rwxr-xr-x 3.0 unx 7791416 bx defN 21-Oct-29 12:23 perccli -rwxrwxr-x 3.0 unx 559164 bx defN 19-Jan-16 21:31 sas2ircu -rwxrwxr-x 3.0 unx 562560 bx defN 19-Jan-16 21:31 sas3ircu 3 files, 8913140 bytes uncompressed, 3238809 bytes compressed: 63.7%
Note: vendor tools may be updated over time. The charm verifies that the shared binaries match a set of known checksums. If you feel a checksum is missing, please file a bug (see link below) and it will be added.
Manufacturer option needs to be left in auto mode.
Please contact the Nagios charmers via the "Submit a bug" link.