Schedules backups not starting, manual or onetime does......

obiwan2

ADSM.ORG Member
Joined
Dec 8, 2005
Messages
184
Reaction score
0
Points
0
Location
Cologne / Germany
Website
Visit site
PREDATAR Control23

Dear Guys,

first for all of you: HAPPY NEW YEAR !!!!

But second.......I have a big big big problem.

Ich have a clustersystem with two nodes, installed with MS Windows Server 2019 Standard.
The TSM-Clients are of version 8.1.11.0
The TSM-Server is version 7.1.7.100 on Redhat Enterprise LINUX

My problem is in short words:

schedules backups are not running
Onetime-Schedule ist running without a problem
Manual backup is running without a problem.
NO entry in dsmerror.log
Only ANR 2578W in actlog of the TSM-Server

This is my dsm.opt:


NODE ZDHPROFILES-UEV
CLUSTERNODE YES
PASSWORDACCESS GENERATE
DOMAIN U:
HTTPPORT 1587
TCPSERVERADDRESS tsm05-backup.dwelle.de
TCPPORT 1500
RESOURCEUTILIZATION 2
TXNBYTELIMIT 5G
TCPWINDOWSIZE 63
TCPNODELAY yes
LANG AMENG
SCHEDMODE PROMPTED


ERRORLOGNAME U:\Shares\profile-uev\TSM\dsmerror.log
SCHEDLOGNAME U:\Shares\profile-uev\TSM\dsmsched.log
ERRORLOGRETENTION 10
SCHEDLOGRETENTION 10

INCLEXCL "U:\Shares\profile-uev\TSM\inclexcl.txt"

snapshotproviderimage VSS
Snapshotproviderfs VSS
domain.image U:


Maybe some of you have already solved this problem ? I'm at the end of my knowledge....

Thanks
Hubert
 
PREDATAR Control23

Dear Guys,

first for all of you: HAPPY NEW YEAR !!!!

But second.......I have a big big big problem.

Ich have a clustersystem with two nodes, installed with MS Windows Server 2019 Standard.
The TSM-Clients are of version 8.1.11.0
The TSM-Server is version 7.1.7.100 on Redhat Enterprise LINUX

My problem is in short words:

schedules backups are not running
Onetime-Schedule ist running without a problem
Manual backup is running without a problem.
NO entry in dsmerror.log
Only ANR 2578W in actlog of the TSM-Server

This is my dsm.opt:


NODE ZDHPROFILES-UEV
CLUSTERNODE YES
PASSWORDACCESS GENERATE
DOMAIN U:
HTTPPORT 1587
TCPSERVERADDRESS tsm05-backup.dwelle.de
TCPPORT 1500
RESOURCEUTILIZATION 2
TXNBYTELIMIT 5G
TCPWINDOWSIZE 63
TCPNODELAY yes
LANG AMENG
SCHEDMODE PROMPTED


ERRORLOGNAME U:\Shares\profile-uev\TSM\dsmerror.log
SCHEDLOGNAME U:\Shares\profile-uev\TSM\dsmsched.log
ERRORLOGRETENTION 10
SCHEDLOGRETENTION 10

INCLEXCL "U:\Shares\profile-uev\TSM\inclexcl.txt"

snapshotproviderimage VSS
Snapshotproviderfs VSS
domain.image U:


Maybe some of you have already solved this problem ? I'm at the end of my knowledge....

Thanks
Hubert

Additional information: Redhat-Version is:
Red Hat Enterprise Linux Server release 6.10 (Santiago)
 
PREDATAR Control23

Are the schedules failed or missed when looking on the server?

I see you are using SCHEDMODE PROMPTED , what does the server activity log show when trying to contact the client to run the schedule?

If you are not using a firewall, you may want to try using SCHEDMODE POLLING.
 
PREDATAR Control23

Hi marclant....here is the excerpt of the actlog....i created a new schedule starting at 13:00:00 for testing.
I get only "missed"...not "failed"
-----------------------------------------------------------------------------------------------------------------
01/05/2021 13:00:01 ANR2578W Schedule TESTSCHED in domain STD_CONT for node
ZDHPROFILES-UEV has missed its scheduled start up window.
01/05/2021 13:00:20 ANR0407I Session 75331 started for administrator ADMIN
(WinNT) (Tcp/Ip tsmmgr(49682)). (SESSION: 75331)
01/05/2021 13:00:21 ANR2017I Administrator ADMIN issued command: select
total_Space_mb,used_space_mb,int((archlog_used_fs_mb /
archlog_tol_fs_mb) * 100),archlog_used_fs_mb from log
(SESSION: 75331)
01/05/2021 13:00:21 ANR0405I Session 75331 ended for administrator ADMIN
(WinNT). (SESSION: 75331)
01/05/2021 13:00:21 ANR0407I Session 75332 started for administrator ADMIN
(WinNT) (Tcp/Ip tsmmgr(49683)). (SESSION: 75332)
01/05/2021 13:00:21 ANR2017I Administrator ADMIN issued command: select
distinct stgpool_name,pct_migr,migr_running,pct_utilized
from stgpools p,devclasses r where
(p.stg_type='DIRECTORY') or ((r.devclass_name=p.devclass)
and (r.devtype='FILE' or p.devclass='DISK')) (SESSION:
75332)
01/05/2021 13:00:21 ANR0405I Session 75332 ended for administrator ADMIN
(WinNT). (SESSION: 75332)
01/05/2021 13:00:21 ANR0407I Session 75333 started for administrator ADMIN
(WinNT) (Tcp/Ip tsmmgr(49684)). (SESSION: 75333)
01/05/2021 13:00:21 ANR2017I Administrator ADMIN issued command: select
count(*) from sessions where session_type='Node'
(SESSION: 75333)
01/05/2021 13:00:21 ANR0405I Session 75333 ended for administrator ADMIN
(WinNT). (SESSION: 75333)
.....
.....
01/05/2021 13:19:05 ANR0406I Session 75346 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59710)). (SESSION: 75346)
01/05/2021 13:19:05 ANR0403I Session 75346 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75346)
01/05/2021 13:19:11 ANR0406I Session 75347 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59713)). (SESSION: 75347)
01/05/2021 13:19:11 ANR0403I Session 75347 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75347)
.....
.....
No further messages in actlog for this client.

I think about testing SCHEDMODE=POLLING also.....I don't know if this would make sense....?
 
PREDATAR Control23

Here a short message version only selected for ZDHPROFILES-UEV:

13:46:30 TSM05 : q act begint='12:50:00' se='ZDHPROFILES-UEV'
Date/Time Message
-------------------- ----------------------------------------------------------
01/05/2021 13:00:01 ANR2578W Schedule TESTSCHED in domain STD_CONT for node
ZDHPROFILES-UEV has missed its scheduled start up window.
01/05/2021 13:19:05 ANR0406I Session 75346 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59710)). (SESSION: 75346)
01/05/2021 13:19:05 ANR0403I Session 75346 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75346)
01/05/2021 13:19:11 ANR0406I Session 75347 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59713)). (SESSION: 75347)
01/05/2021 13:19:11 ANR0403I Session 75347 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75347)
01/05/2021 13:20:24 ANR0406I Session 75354 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59742)). (SESSION: 75354)
01/05/2021 13:20:25 ANR0403I Session 75354 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75354)
01/05/2021 13:20:31 ANR0406I Session 75355 started for node ZDHPROFILES-UEV
(WinNT) (Tcp/Ip 172.30.0.126(59746)). (SESSION: 75355)
01/05/2021 13:20:31 ANR0403I Session 75355 ended for node ZDHPROFILES-UEV
(WinNT). (SESSION: 75355)
01/05/2021 13:46:30 ANR2017I Administrator SERVICE.HEIDERH issued command:
QUERY ACTLOG begint=12:50:00 se=ZDHPROFILES-UEV
(SESSION: 75370)

One schedule should have started at 13:00:00, after changing at 13:30:00
But nothing happened....

I checked the client-IP address with pinging it from the server... this is working. The other way (client to server) also working...
 
PREDATAR Control23

I see you are using
Should I try an older version of the client ?
No, try polling instead. Prompted is not as robust as polling, should only be used in 2 specific situation:
- client is behind a firewall and can only be contacted by the server
- there's a requirement to run ad hoc schedules or clientactions immidiately

With prompted, if the client is not contacted by the server, the issue is reported on the server, you have to dig further in the activity log. You will see this message when the server tries to contact the client: ANR2561I Schedule prompter contacting node name (session session number) to start a scheduled operation., and in the messages that follow, you will likely see a failure with the reason. Usually ANR2716E or ANR2717E, but could be something else.
 
Last edited:
PREDATAR Control23

Sometimes, using polling once, then switching back to prompted works. Somehow in polling, it updates the client definition on the server. But unless you have a requirement for prompted (the 2 I mentioned earlier), you will be better served with polling, it's much more reliable.
 
PREDATAR Control23

Good morning marclant,

I have now tested and observed the behavior of individual cluster resources.

On Node1, the backups only start in POLLING mode
On Node2, the backups start in POLLING and also in PROMPTED mode

Strange, isn't it ?

Question: Is it a problem for the cluster-functionality itself if running the backups in POLLING mode ?

Thank you very much in advance !
Very best regards

Hubert
 
PREDATAR Control23

Question: Is it a problem for the cluster-functionality itself if running the backups in POLLING mode ?
No issues, that's the default.

Strange, isn't it ?
No really. Polling has a lot less dependencies and easier to get going. I wouldn't be surprised if you switched node1 to prompted, it would now work. However, given polling is more robust, I'm not sure why you would want to do that.
 
Top