topic Re: R80.40 - can't cancel task in API / CLI Discussion

R80.40 - can't cancel task

John_Fleming — Wed, 23 Dec 2020 20:54:07 GMT

I started using the python SDK to set log servers. In my first attempt I was playing with setting the log server to a CLM for a given simple cluster. I noticed afterwards I didn't have anything in the API call to publish so I went to the gui and hit publish and disconnect now its just sitting at %10.

Any idea how to cancel this task? Discard & disconnect rows "An internal error occurred."

Re: R80.40 - can't cancel task

John_Fleming — Wed, 23 Dec 2020 21:38:40 GMT

So looks like setting the log server via API will basically eff up the API on the CMA until I mdsstop. 😕

This is JH 87 BTW.

Re: R80.40 - can't cancel task

PhoneBoy — Thu, 24 Dec 2020 01:57:18 GMT

Ouch, sounds like a TAC case is in order.
Also tagging @Omer_Kleinstern

Re: R80.40 - can't cancel task

John_Fleming — Thu, 24 Dec 2020 04:26:33 GMT

I haven't had a chance to debug or bother diamond, which i'll wait until next week.

Re: R80.40 - can't cancel task

Alex_Zelinsky — Thu, 24 Dec 2020 09:33:40 GMT

Please specify more details, is this set-simple-cluster api? And what exactly did you try to do

Re: R80.40 - can't cancel task

John_Fleming — Thu, 24 Dec 2020 17:31:47 GMT

I plan on creating a replication I just haven't had time. Yes set simple cluster and send-logs-to-server. I set name and send-logs-to-server as a single member list.

The docs say its an object but others say string or list and description says Servers so i'm not %100 what I should be sending. I didn't try logs and alerts (or whatever that other one was).

Re: R80.40 - can't cancel task

Alex_Zelinsky — Sun, 27 Dec 2020 06:24:53 GMT

I set it as follows:

Here is an example of setting log server for a cluster:
mgmt_cli --session-id "$session_id" set-simple-cluster name "$CLUSTERNAME" ipv4-address "$CLUSTERMGMTIP" ipv6-address "$IPV6_CLUSTERMGMTIP" \
	version "$GWVERSION" os-name "Gaia" cluster-mode "opsec-ha" firewall true ips false vpn false \
	send-logs-to-server.0 "$MGMTNAME" send-alerts-to-server.0 "$MGMTNAME"
It's network object name, and you can provide a few.
However the issue of cancelling task is not related.
Actually there is now way to cancel a task.
To better understand your issue I need more details and cpm.elg when you're getting INteranl Error.
I undertsatand you get that error when you try to discradand disconnect API session from SC, right?

Re: R80.40 - can't cancel task

John_Fleming — Mon, 28 Dec 2020 16:20:46 GMT

Hi Alex, I haven't had a chance to put something meaningful in here yet. I can't send logs or anything as the customer will not allow this. Its a pretty strict env. I'll see if I can replicate and send is the smallest replication of the python code to trigger when I get a few. Should be this week.

Re: R80.40 - can't cancel task

John_Fleming — Mon, 28 Dec 2020 23:33:32 GMT

Test env is stood up. I'll start working on the code python code next.

First.. Nachos for they indeed rule.

Re: R80.40 - can't cancel task

John_Fleming — Tue, 29 Dec 2020 01:57:20 GMT

Eff.. couldn't replicate. I'm going to move this code back to the lab and see if it fails again. I'll check log files you pointed out and engage support if needed.

Any input on debugs I should turn on?

Re: R80.40 - can't cancel task

John_Fleming — Tue, 29 Dec 2020 02:02:08 GMT

Here is the python code.

from cpapi import APIClient, APIClientArgs import pprint pp = pprint.PrettyPrinter(indent=4) api_server='10.1.18.100' username='admin' password='somepw' domain_name='Test_Domain' def main(): client_args = APIClientArgs(server=api_server) with APIClient(client_args) as domain_client: login_res = domain_client.login(username, password, domain=domain_name) if login_res.success is False: print("Try upgrading to R81 maybe?") exit(81) json_data = { "name": "TestCluster", "send-logs-to-server": "TestDomainLogServer" } print("Calling set cluster api") simple_cluster_output = domain_client.api_call("set-simple-cluster", json_data) pp.pprint(simple_cluster_output) print("Calling publish api") publish_output = domain_client.api_call("publish",{}) pp.pprint(publish_output) if __name__ == "__main__": main()

Re: R80.40 - can't cancel task

Alex_Zelinsky — Tue, 29 Dec 2020 07:08:59 GMT

turn on the following logs in $MDS_FWDIR/conf/tdlog.com

log4j.logger.com.checkpoint.management.gateway_legacy.services.internal.LegacyGatewayObjectCrudSvcImpl=DEBUG
log4j.logger.com.checkpoint.management.gateway_legacy.objects.cluster.ClusterUtils=DEBUG
log4j.logger.com.checkpoint.management.gateway_legacy.web_services.internal.LegacyGatewayObjectCrudSvcRemoteImpl=DEBUG

What version do you run? R80.40 jumbo?

Re: R80.40 - can't cancel task

Alex_Zelinsky — Tue, 29 Dec 2020 07:10:47 GMT

Python code seems pretty simple and straightforward.

Do you see the following line printing:

pp.pprint(simple_cluster_output)

and what does it print?

Re: R80.40 - can't cancel task

John_Fleming — Tue, 29 Dec 2020 15:39:54 GMT

Everything ran clean on the replication so nothing useful other then http 200 etc.

Re: R80.40 - can't cancel task

Alex_Zelinsky — Wed, 30 Dec 2020 08:13:19 GMT

well, let me know if it happens again