[oi-dev] hotplug errors with scsi-vhci after update

[Udo Grabowski (IMK)]

On 08/11/2018 17:07, Udo Grabowski (IMK) wrote:
> Hi, upgraded to the latest Hipster from Sep-13 to today, and got a massive
> amount of error messages from hotplug, for each of the 70 (SAS-expander
> connected) disks I got the block of messages below (in total 2325 lines!).
> But the pools are available and healthy.
>
> The disks are configured in /kernel/drv/scsi_vhci.conf:
> scsi-vhci-failover-override=
>         "HGST    HUSMH8010BSS200", "f_sym",
>         "HGST    HUS726040AL5210", "f_sym";
> device-type-scsi-options-list=
>         "HGST    HUSMH8010BSS200", "symmetric-option",
>         "HGST    HUS726040AL5210", "symmetric-option";
> and are 70 two-port disks, crosswise visible to a second host in the
> same machine, 35 disks per host configured into a pool per host,
> usually the unused disks are unconfigured on the resp.second host (but
> see problem below). The messages are for all disks, regradless if they
> belong to the host or not, and have not been thrown before the upgrade
> (even though the next problem already was visible).
>
> For quite a while now, I noticed another, unrelated error: I cannot
> unconfigure any disk with 'cfgadm -c unconfigure', it's immediatly
> reconfigured (regardless of hotplugd running or not). This is awkward
> especially when a disk is behaving bad.
>
> Any ideas what's wrong here ?
> ....
> Nov  8 16:43:12 imksunth7 hald[268]: [ID 702911 daemon.error] 16:43:12.794 [E]
> hotplug.c:66: devpath /scsi_vhci/disk at g5000cca2441ed3f4 already present in
> store, ignore event
> Nov  8 16:43:12 imksunth7 hald[268]: [ID 702911 daemon.error] 16:43:12.795 [E]
> hotplug.c:66: devpath /scsi_vhci/disk at g5000cca2441ed3f4:wd already present in
> store, ignore event
> Nov  8 16:43:12 imksunth7 hald[268]: [ID 702911 daemon.error] 16:43:12.795 [E]
> hotplug.c:81: Parent is NULL devfs_path=/scsi_vhci/disk at g5000cca2441ed3f4:wd/d0
> parent_udi=/org/freedesktop/Hal/devices/scsi_vhci_0/disk180_0/sd180
//openindiana.org/mailman/listinfo/oi-dev
>

Found the culprit, enabled --verbose and --use-syslog in HAL service
script a while ago for a test, and didn't revert it. Probably the
cause for the excessive messages.

 > For quite a while now, I noticed another, unrelated error: I cannot
 > unconfigure any disk with 'cfgadm -c unconfigure', it's immediatly
 > reconfigured (regardless of hotplugd running or not). This is awkward
 > especially when a disk is behaving bad.

Neverthelesss, this problem is still present, and I don't find a
reasonable cause for it. It started when upgrading from an older
Hipster from 2016 to the 2018.4 snapshot, so a wider range of commits
to illumos are to be suspected here.
-- 
Dr.Udo Grabowski   Inst.f.Meteorology & Climate Research IMK-ASF-SAT
http://www.imk-asf.kit.edu/english/sat.php
KIT - Karlsruhe Institute of Technology           http://www.kit.edu
Postfach 3640,76021 Karlsruhe,Germany T:(+49)721 608-26026 F:-926026

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5227 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://openindiana.org/pipermail/oi-dev/attachments/20181109/4ccd8df6/attachment-0005.bin>