Virtual Switches in VSX cluster are shown in "PROBLEM" status in SmartView Monitor without any error message

Support Center > Search Results > SecureKnowledge Details

Technical Level

Solution ID	sk112067
Technical Level
Product	VSX, SmartView Monitor
Version	R77.30 (EOL), R80.10
OS	Gaia
Date Created	22-Jun-2016
Last Modified	02-Oct-2018

Symptoms

SmartView Monitor shows "PROBLEM" status for Virtual Switches objects of a VSX cluster without any error message.
There is no impact on traffic passing through these Virtual Switches.

Output of the "cpwd_admin list -ctx <VSID>" command on VSX cluster members shows that FWD daemon is not running (the "STAT" column shows "T" - i.e., terminated).

Example:

[Expert@vsx2:0]# cpwd_admin list -ctx 3
APP        CTX        PID    STAT  #START  START_TIME             MON  COMMAND
FWK_WD     1          8512   E     1       [11:13:04] 7/6/2016    N    fwk_wd -i 1 -i6 0
CPD        1          8728   E     1       [11:13:05] 7/6/2016    Y    cpd
FWD        1          0      T     3       [11:15:15] 7/6/2016    N    fwd
CPVIEWD    1          8934   E     1       [11:13:07] 7/6/2016    N    cpviewd

Starting the FWD daemon on VSX cluster members manually under debug shows that registration of the Critical Device (pnote) for FWD daemon failed:
[Expert@HostName:0]# fwd -d ... ... [ PID ...]@VSX_HostName[Date Time] fwd_monitor_init: fwd_monitor_active was not defined or set to 0, we will not register to active monitoring [ PID ...]@VSX_HostName[Date Time] cpuser_socket_connect: Connecting to server /vs1/dev/fw0 for /dev/fw0 ... ... Error while opening pnote_init_retries.conf file. [ PID ...]@VSX_HostName[Date Time] fwha_drv_initialize_pnote: registering pnote fwd [ PID ...]@VSX_HostName[Date Time] cpuser_socket_connect: Connecting to server /vs1/dev/fw_fwhapnote0 for /dev/fw_fwhapnote0 ... ... [ PID ...]@VSX_HostName[Date Time] fwha_drv_initialize_pnote: error ioctl /dev/fw_fwhapnote0 Failed to register pnote
Issue occurs only in VSX cluster running R77.30 with R77.30 Jumbo Hotfix Accumulator - Take 162 and above.

Cause

R77.30 Jumbo Hotfix Accumulator (since Take_143) contains a kernel parameter "fwha_installed_drv_initialize_pnote_process_err"
that is involved in checking whether Cluster Layer is already enabled on the cluster member before registering a Critical Device (Pnote) - refer to sk110073.
By default, this kernel parameter is enabled (its value is set to "1").

Due to incorrect check in Take_162 and above, this check fails on Virtual Switches.
As a result, registration of the Critical Device (pnote) for FWD daemon fails on Virtual Switches.

Solution

Note: To view this solution you need to Sign In .