12-08-2010 08:49 AM
When I try to bring online Sybase ASE within 2 minutes after ASE brings up I get the error
VCS ERROR V-16-2-13066 Thread(3) Agent is calling clean for resource(Syb_Sybase) because the resource is not up even after online completed.
And agent kills the ASE. Right before ASE gets killed I can connect to it.
So the agent can successfully start and shutdown ASE (using isql), but can't understand than after start ASE is working normally.
The question is why agent can't figure out that ASE is ONLINE and works correctly?
Solved! Go to Solution.
12-09-2010 12:58 AM
I have faced a similar issue on sybase database and it ended up because of the fact that vcs monitoring uses the ps -ef command to validate the instance name.
Request DBA to modify the RUN file in such a way that -s<instance> is the first argument..I guess in your case it is specified at the last.
Thanks
12-08-2010 09:23 AM
Is the issue that VCS does not detect the ASE is online at all, or is it that the ASE is taking a longer time to come up than VCS is expecting (timing issue)?
If it's the latter, can you try increasing OnlineWaitLimit for the Sybase resource type so it waits longer for before deciding the ASE is offline? (the default is 2, so since you said that you can connect to it right before it's shutdown, perhaps try increasing to 4 so it tries 4 times instead of twice?)
See VCS User's Guide -> Appendixes -> VCS attributes -> Resource type attributes:
http://sfdoccentral.symantec.com/sf/5.0MP3/solaris/html/vcs_users/ap_vcs_attributes4.html
--------------------
OnlineWaitLimit (user-defined)
Note: This attribute can be overridden.
Number of monitor intervals to wait for the resource to come online after completing the online procedure. Increase the value of this attribute if the resource is likely to take a longer time to come online.
* Type and dimension: integer-scalar
* Default: 2
--------------------
If you have tried this and the problem is that the agent never detects the ASE, then please provide additional details to investigate, at minimum:
- OS/platform
- VCS version
- VCS configuration (main.cf)
hope that helps,
Grace
12-08-2010 08:10 PM
Hi,
Have you tried onlining this ASE outside VCS ? If you can try this, it will clearly give you indications on how much time it takes to come online. Once you have the exact timelines, you can define the parameters accordingly in VCS.
I am assuming that every thing is working as expected outside to VCS.
In addition to what Grace suggested, you can also tweak around OnlineTimeout, tolerencelimit etc.
Have a look around VCS user's guide to know more about these tunables. You can find the user's guide on below link, check for your OS & veritas version
http://sort.symantec.com/documents
Gaurav
12-08-2010 10:44 PM
Hi,
Have you configured Detailed Monitoring on the Sybase Resource? If so, please set it to 0, and retry the monitor operation. If you find that the resource probe show that it is now online, please check your detailed monitoring configuration.
It could be that because you've set detailed monitoring, VCS is not able to access the table, or the user account isn't granted access, etc, etc.
12-09-2010 12:11 AM
Hi,
Yes, I've configured detailed monitoring but disabling or enabling it doesn't change anything.
12-09-2010 12:16 AM
Can you post the log for sybase from /var/VRTSvcs/log/ here please.
12-09-2010 12:58 AM
I have faced a similar issue on sybase database and it ended up because of the fact that vcs monitoring uses the ps -ef command to validate the instance name.
Request DBA to modify the RUN file in such a way that -s<instance> is the first argument..I guess in your case it is specified at the last.
Thanks
12-09-2010 01:45 AM
Hi
OS: Sun Solaris 10 x86 on Vmware ESX 4.0
VCS version: 5.1
Sybase ASE 15.5
VCS config:
include "types.cf"
include "SybaseTypes.cf"
cluster cl1 (
UserNames = { admin = bQNkNMmNNnNP, testuser = hifGfgEffDfd }
ClusterAddress = "192.168.1.52"
Administrators = { admin, testuser }
)
system sol10_cl1_1 (
)
system sol10_cl1_2 (
)
group SybaseASE (
SystemList = { sol10_cl1_2 = 1, sol10_cl1_1 = 0 }
)
DiskGroup Syb_DiskGroup_Data (
DiskGroup = dg2
)
IP Syb_IP (
Device = e1000g0
Address = "192.168.1.52"
NetMask = "255.255.255.0"
)
Mount DataMount (
MountPoint = "/mnt/shdsk1"
BlockDevice = "/dev/vx/dsk/dg2/volume1"
FSType = vxfs
FsckOpt = "-y"
)
NIC Syb_NIC (
Enabled = 0
Device = e1000g0
)
Sybase Syb_Sybase (
Server = SOL10_CL1
Owner = sybase
Home = "/opt/Sybase/ase_155"
Version = "15.5"
SA = sa
SApswd = JPKpLPmPNpOPp
DetailMonitor = 1
User = sysmon
UPword = cqfQlqFopOroQ
Db = mondb
Table = montbl
WaitForRecovery = 1
OnlineTimeout = 420
OnlineWaitLimit = 4
ToleranceLimit = 2
)
DataMount requires Syb_DiskGroup_Data
Syb_IP requires Syb_NIC
Syb_Sybase requires DataMount
Syb_Sybase requires Syb_IP
// resource dependency tree
//
// group SybaseASE
// {
// Sybase Syb_Sybase
// {
// Mount DataMount
// {
// DiskGroup Syb_DiskGroup_Data
// }
// IP Syb_IP
// {
// NIC Syb_NIC
// }
// }
// }
Still the agent can't detect that ASS is online and shortly after ASE start shutdowns it. By the way how does the agent check if ASE is online?
12-09-2010 01:55 AM
Thank you very very much for this tip. It indeed solves the problem.
I think that Veritas Cluster Server Agent for Sybase Installation and Configuration Guide must have that request to rewrite RUN script