Natural sorting of SONiC config gen output consumes lot of CPU cycles.
The sole use of natsorted was to make test comparison easier and so,
the natsorting logic is now relocated to the test suite. As a result
sonic-cfggen gained nearly 1 sec per call since we no longer import
natsorted module!
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Argument to write to config-db is not allowed when using template.
This PR allows cfggen to write to redis db when using template
mode.
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to cfggen take considerable time. With batch mode, we will have the ability
to reduce number of calls from services.
Example of the batch mode command:
sonic-cfggen -t template-1.j2 -t template-2.j2,config-db -t template-3.j2,config-db -t template-4.j2,file1 -t template-5.j2,file2 --write-to-db.
template-1.j2 will be rendered to stdout since it is missing the dest part. stdout is default
config-db is a special keyword that will inject the rendered template into internal data structure. The internal data structure gets written to redis-db with --write-to-db switch. In the case the user would like to write to a file named config-db, it could be given as /config-db or ./config-db
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Optimizing number of calls made to sonic-cfggen during service
start up as it adds to total system boot up time.
***-Test 1***
there is an average saving of 1 to 1.5 sec between old script and new script
```
root@str-s6000-acs-14:/# time /usr/bin/rest-server-old.sh
Generating temporary TLS server certificate ...
2020/07/09 19:03:33 wrote cert.pem
2020/07/09 19:03:33 wrote key.pem
REST_SERVER_ARGS = -ui /rest_ui -logtostderr -cert /tmp/cert.pem -key /tmp/key.pem
/usr/sbin/rest_server -ui /rest_ui -logtostderr -cert /tmp/cert.pem -key /tmp/key.pem
real 0m8.790s
user 0m7.993s
sys 0m0.584s
root@str-s6000-acs-14:/# time /usr/bin/rest-server-new.sh
Generating temporary TLS server certificate ...
2020/07/09 19:03:45 wrote cert.pem
2020/07/09 19:03:45 wrote key.pem
REST_SERVER_ARGS = -ui /rest_ui -logtostderr -cert /tmp/cert.pem -key /tmp/key.pem
/usr/sbin/rest_server -ui /rest_ui -logtostderr -cert /tmp/cert.pem -key /tmp/key.pem
real 0m6.940s
user 0m5.670s
sys 0m0.386s
```
***-Test 2***
Built an image with this change and rest server is running with params as described in test 1 above
```
admin@str-s6000-acs-14:~$ ps -ef | grep rest_server
root 3301 2866 2 02:09 pts/0 00:00:10 /usr/sbin/rest_server -ui /rest_ui -logtostderr -cert /tmp/cert.pem -key /tmp/key.pem
```
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to one call during startup when starting swss service.
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to two calls during startup when starting frr service.
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to one call during startup when starting radv service.
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to one call during startup when starting dhcp-relay
service.
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to one call during startup when running interfaces-
config.
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Calls to sonic-cfggen is CPU expensive. This PR reduces calls to
sonic-cfggen to once calle during snmp startup
singed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
This PR enables cfggen to readr/write from Redis DB using pipelines.
Pipelines enables batch read/write from/to Redis DB.
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
This PR limited the number of calls to sonic-cfggen to one call
per iteration instead of current 3 calls per iteration.
The PR also installs jq on host for future scripts if needed.
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Support ACL Table type Mirrorv6 for Innovium (#1528)
Enable v6 ACL rule based Mirroring for Innovium Platform
Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>
To limit IO and space usage on the flash device the boot0 script makes sure the SWI is in memory.
Because SONiC maps /tmp on the flash, some logic is required to make sure of it.
However it is possible for some provisioning mechanism to already download the swi in a memory file system.
This was not properly handled by the boot0 script.
It now properly detect if the image is on a tmpfs or a ramfs and keep it there if that is the case.
- How I did it
Check the filesystem on which the SWI pointed by swipath lies.
If this filesystem is a ramfs or a tmpfs the move_swi_to_tmpfs becomes a no-op.
Made sure the cleanup logic would not behave unexpectedly.
- How to verify it
In SONiC:
Download the swi under /tmp and makes sure it gets moved to /tmp/tmp-swi which gets mounted for that purpose.
Make sure /tmp/tmp-swi gets unmounted once the install process is done.
Create a new mountpoint under /ram using either ramfs or tmpfs and download the swi there.
Install the swi using sonic-installer and makes sure the image doesn't get moved by looking at the logs.
* Parse device type from <ElementType> first in <PngDec>
* Fall back to <Device> type attribute if no <ElementType> is found
Signed-off-by: Lawrence Lee <lawlee@microsoft.com>
- Why I did it
The change is done to make sure the system initialization is done before the hostcfgd sets the feature states.
- How I did it
This is port of the PR #6232.
Since the systemctl version in 201911 doesn't support "--wait".
Added a function to check the output of systemctl is-system-running every second, till the command system is done booting up.
For now this change is only applicable to multi asic platforms based on the testing this change will be extended to all platforms in the future PR.
Signed-off-by: Arvindsrinivasan Lakshmi Narasimhan <arlakshm@microsoft.com>
Fixes#5663
- Why I did it
It's currently possible for the SNMP timer to conflict with config reload (specifically if the timer triggers while config reload is stopping the SWSS service). config reload triggers SWSS to shutdown, which causes SNMP to shutdown, which conflicts with the SNMP timer causing SNMP to startup. See the linked issue for more details.
- How I did it
Including the After ordering dependency forces the SNMP timer to wait until SWSS finishes stopping, preventing the conflict. If there is an ordering dependency between two units (e.g. one unit is ordered After another), if one unit is shutting down while the other is starting up, the shutdown will always be ordered before the startup. In this case, that means that the SNMP timer is forced to wait for the SWSS shutdown to complete. Only then can the SNMP timer proceed. See here for more details.
It's important to note that the After dependency will not cause SWSS to be started when the SNMP timer fires (assuming that SWSS has not yet been started). The existing Requisite dependency in the SNMP service will also not cause SWSS to be started, instead it will cause the SNMP service to fail if SWSS is not active.
Signed-off-by: Lawrence Lee <lawlee@microsoft.com>
- Why I did it
The l2switch.j2 template does not include all fields for PORT. This could be incompatible with the 201911 image or later.
- How I did it
Update l2switch.j2 template and add a unit test.
In order to install a SONiC image on top of a NVMe SSD disc properly with ONIE we must configure it properly on the installer.sh script.
Signed-off-by: Shlomi Bitton <shlomibi@nvidia.com>
b909766aab63da5e9a51e05fd2bf79e80db75e5 (HEAD -> 201911, origin/201911) Fix show ip/v6 route summary non-multi-asic platform to interact with FRR directly (#1306)
057d2ee26586034975e21a5cacb1a00ca87f2857 Add support to collect tech support on multi ASIC platform (#1308)
38ab16d5835b917f7459044853276c9d4b53c98b [CLI][PFCWD] Fix issue with specifying ports in pfcwd start on masic platforms (#1203)
Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>
7f50b9815e14d90c02d9dce63fd08d90e25cee3f (HEAD -> 201911, origin/201911) handled update() function of fdb orchagent for FDB FLUSH event (#1534)
17adc13b6ca21846fe27c94d6a16f9909c712d77 Add a check for warm-restart, and do a clear only when warm-restart is enable. (#1498)
d097260a5aa7bd611babd5062e220056374e23d8 Fixed compilation failure with debug option (#1518)
Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>
Usually for a use case like networking - should not be configured to reach c6, the maximum used is c1e – due to the added latency getting in & out of states (bad for networking).
Following a recommendation by Intel, networking system should avoid getting in & out of states which introduce latency. The recommended state is c1e and no state change enabling.
In addition, c-state sole purpose is to save power and when inside a networking switch its really negligent being such a tiny consumer vs. the whole cluster.
Signed-off-by: Shlomi Bitton <shlomibi@nvidia.com>
EEPROM cache file is not refreshed after install a new ONIE version even if the eeprom data is updated. The current Eeprom class always try to read from the cache file when the file exists. The PR is aimed to fix it.
The issue was a typo introduced in #6006. In that change, the BGP allow list
configuration manager was updated to use a method of common ConfigMgr
for restarting peer groups. However, the method name 'restart_peers' was
used instead of the correct 'restart_peer_groups'.
This change updated the managers_allow_list.py to use correct method
'restart_peer_groups' for restarting peer groups.
Signed-off-by: Xin Wang <xiwang5@microsoft.com>
- Why I did it
Fix issue #6043
- How I did it
We are disabling in container frr log. The log entries are sent to base image and are logged in /var/log/quagga/bgpd.log.
However, we need to remove the whole outchannel config block to avoid an error message raised by rsyslogd.
- How to verify it
Without the change, test_autorestart bgp container will fail on loganalyer errors. With the change, restarting bgp container is no longer generating error message and the test will pass.
The log generated by frr continued appearing in /var/log/quagga/bgpd.log
* [bgpcfgd]: Batch bgp updates.
vtysh -f command is slow. It is sometimes takes about 3 seconds.
When we need to run many vtysh -f commands that slows down the system.
Batch vtysh -f updates.
* Use correct file to import run_command
Added new flag value 'always_enabled' for the state and auto-restart field of feature table
init_cfg.json is updated to initialize state field of database/swss/syncd/teamd feature and auto-restart field of database feature
as always_enabled
Once the state/auto-restart value is initialized as "always_enabled" it is immutable and cannot be change via feature config commands. (config feature..) PR#Azure/sonic-utilities#1271
hostcfgd will not take any action if state field value is 'always_enabled'
Since we have always_enabled field for auto-restart updated supervisor-proc-exit-listener
not to have special check for database and always rely on value from Feature table.
