SQL 2012 :: Native Backups Hanging Across Instances
Jan 12, 2015
Native backups to NAS do not complete.We have been experiencing an issue whereby our native backups are hanging with status': SUSPENDED/ RUNNABLE...I ran select * from sys.sysprocesses. All of the backup SPIDs processes show BACKUPTHREAD/PREEMPTIVE_OS_FILEOPS
This first occurred last Wednesday evening. When I discovered this on Thursday, I attempted to kill the backup jobs. This also hung with 0% completed/0% time remaining. Backups hung on more than one instance.That evening, I attempted to restart the instance which also failed with something along the lines of: could not start MASTER file in use.
I then restarted the server--which I really did not want to do--and this cleared it. I was also able to manually kick off maintenance plans (DBCC CHECKDB and full backup) without issue.I was off Friday and the weekend. I came in this morning and found the maintenance plans (diff/tlog backups) did not complete in some of the instances--in one case, the instance affected now was not affected before. They appeared to have hung on their next scheduled kickoff which was later that night after I went home
Remembering the "file in use" error, I have run process monitor to see if anything unusual had a lock on any files. I saw only SQL Server and Double-Take processes accessing log files.Being a relatively new DBA, I am user where to go next in trying to track down the cause of this issue. This is fairly urgent as one of the instances that has had this problem both times is our production SharePoint environment.
ENVIRONMENT:
SQL version:
Microsoft SQL Server 2012 (SP1) - 11.0.3368.0 (X64)
May 22 2013 17:10:44
Copyright (c) Microsoft Corporation
Enterprise Edition: Core-based Licensing (64-bit) on Windows NT 6.2 <X64> (Build 9200: )
Can we backup our cluster databases directly to tape using native backups (without using any third party tool) ? It's SQL Server 2012 two node Active/Passive cluster. One of the DB will be huge in size, hence checking if we can directly backup from the cluster instance to a tape.
Using merge replication + web synchronization, I have a situation when there are large amount of data changes to upload to the publisher, Merge agent would create a large request and send it over. The publisher gets it and is able to work on it. After few minutes it has finished but (I assume) the connection has been dropped. At the subscriber's side, it appears that the merge agent is hung. The output would look like something like this:
Upload request size is XXX bytes. Uploaded a total of 100 chunks. Uploaded a total of 200 chunks. Uploaded a total of 211 chunks.
The request message was sent to [URL] ....
Normally, when the publisher finishes working, the merge agent then continues processing. But when it takes more than few minutes (it seems to break about at 2 or 3 minutes), merge agent will hang as long as the InternetTimeout setting is (currently 20 minutes) before finally failing and retrying.
But that's not right. The publisher was done and can't communicate back to the merge agent (presumably because the connection was dropped). As a result, merge agent will try to re-enumerate changes on top of giving appearance that it's hung.
I've already fiddled with settings such as MaxUploadChanges, UploadGenerationsPerBatch, UploadReadChangesPerBatch, and UploadWriteChangesPerBatch. However, none of those setting actually ensure that the request message is too large. It has worked in breaking up the changes into separate batches (e.g. processing a single table rather than all tables) which results in more frequent updates and thus avoid the problem.
However, when a single table has several changes, it is still lumped into one large request which then takes more than 2-3 minutes to process on publisher's side and thus I still end up with the same symptom of merge agent hanging.
Is there anything else I could try to get merge agent to keep its connection alive even during processing a large request?
I'm wondering if I can take native SQL log backup in addition to DPM backups. We're using DPM for both databases (daily) and logs (every 15min). So, if I take SQL log backup separately, will it break anything in DPM and we won't be able to perform point it time recovery?
I'm currently trying to using this bcp script to export results to a csv.
USE crm001_spark GO EXEC xp_cmdshell 'bcp "SELECT top 10 * FROM [datebasename].dbo.table name]" queryout "C:Customers.csv" -c -b 10000 -t , -T -S [server]datebasename'
and the ERROR that is coming out is
SQLState = 08001, NativeError = -1 NULL Error = [Microsoft][SQL Native Client]Login timeout expired Error = [Microsoft][SQL Native Client]SQL Network Interfaces: Error Locating Server/Instance Specified [xFFFFFFFF]. SQLState = 08001, NativeError = -1 Error = [Microsoft][SQL Native Client]An error has occurred while establishing a connection to the server. When connecting to SQL Server 2005, this failure may be caused by the fact that under the default settings SQL Server does not allow remote connections. SQLState = S1T00, NativeError = 0
I have checked the server config manager TCP port and it is pointing to the correct port 1433. What else could be causing the error ?
My issue is that: here have two datamirror servers, and when the first fail, mirror needs to assume with the same name.
I'll try explain: SD01 - Primary, SD02 - mirror. If SD01 fails, SD02 assumes called SD01. By the way, the instances are in same domain, but not in same computer.
In our packages, we are using Microsoft SQL Server Native Client 10.0 provider to connect our SQL Server 2008 DBs, which is working fine till we are migrating 2008 DB server to 2012. after the upgrade, some of our packages are working fine, some are not. I'm just curiuos to know will the old provider(Microsoft SQL Server Native Client 10.0) will work fine for connecting 2012 DBs? or I need to update all our packages to re-point to new 2012 Provider.
Is it possible for a clustered instance of SQL2012 to have 2 network names ?
Reason: I need to segregate admin access to a clustered instance so that the admins and SSMS connect via a different IP address than the application. I know I can block SSMS access via application-level firewalls, but ideally the application would connect to CLUSTER1INSTANCE01 on , say, 10.192.5.5, and the admins would connect to CLUSTER1ADMININSTANCE01 on 172.168.2.2, and they'd be the same instance, just using different names and IPs
I am using robocopy to copy my sql server backups to a fileshare. This is working great except for named instances.I use Ola Hallegren's backups which create a folder called xxxx$xxxx. Robocopy can't seem to handle the $.
I am using sql2012 with partitiondb custom installation over 6 dbs on 4 servers, ~200GB per db. I am looking for the perfect graphic tool (similar to Perfmon+Activity Monitor) where I can monitor a wide-scope SQL environment. I am interested in data / log / table / index growth, buffer cache hit ratio, average wait time, physical/logical reads/writes and such. I am interested in real-time / time-range metrics. I know I can issue immediate queries against dm_os_performance_counters / dm_os_buffer_descriptors / dm_exec_query_stats and get some of the relevant data, but is this the only way?Also, for your opinnion, what are the most crucial metrics to monitor when dealing with multiple dbs?
I have inherited a new SQL Server 2008 database server and can not figure out how my user databases are being backed up. This database server is running under a VM.
All the user databases are being backed up nightly per the SQL server log. The backups are written to a virtual disk and is kicked off by the NT AUTHORITYSYSTEM user. I can not see the virtual disk. A restore task does not provide any information about the last backup. I have created a new database, and it is automatically included in the next set of backups.
I have looked at the windows event viewer with out any luck. There are no SQL Server Maintenance Plans or Agent jobs that call a backup. I have also checked the Windows Task Scheduler and can not find any task that does a backup.Could the backups be called from another server ?
I am using Ola Hallegren's scripts to do backups. He uses @Cleanup Time to delete backups older than a certain number of hours. My situation is I do a full backup of a database on Sunday and then I have a few Differentials and then log backups for the rest of the week. When Sunday rolls around again and my full backup is finished, I would like to delete all the differential backups and log backups. Any way that I could accomplish this using Ola's scripts?
I have tried doing sql server backups to a file share and that has been taking too long. So I've decided to backup locally and then taking those backups and getting them off the server. For those that are doing this what do you use to get your backups off the server?
I am sure I have seen in the past in a monitoring tool that PLE drops off to 0 whenever we do a backup. I was doing some reading around this however and found something that said backups use a different portion of memory external to the buffer pool (minmax settings).
Is this correct and how can I tell how much memory will be required for a backup?
I just completed a copy-Only compressed backup of a DB (with a FULL Recovery Model ) on SQL Server 2012 and the resulting backup (the bak file) is 1/100th the size of the data & log file. Is the compression in SQL Server 2012 just that good or did something else happen that I did not catch? Below is the T-SQL to re-create the backup. The size of the data file is 750MB and the log file is 75GB and is %95 used according to the SQLPERF command.
Does the compression in SQL Server 2012 simply that good
BACKUP DATABASE [MYBIGOLEDB] TO DISK = N'Z:Microsoft SQL ServerMSSQL11.MSSQLSERVERMSSQLBackupMYBIGOLEDB_20150611.bak' WITH COPY_ONLY, NOFORMAT, INIT, NAME = N'MYBIGOLEDB-Full Database Backup', SKIP, NOREWIND, NOUNLOAD, COMPRESSION, STATS = 10 GO
So I'm testing some things in our new servers and was trying to restore a database from some striped backup sets. We have 4 files for our backups and restoring the FULL backups with no recovery worked beautifully via SSMS. But when I tried to restore the differentials (also striped across 4 files), the GUI gave me this error:
Unable to create restore plan due to break in the LSN chain.
How to locate when the break happened and I came across this article about how this is an SSMS 2012 bug.
So I tried the advice in this article to attempt a restore via Files and Filegroups, and ended up with the below error:
EDIT: Picture is attached if it is not showing in post.
I was able to restore via T-SQL, but I want to also be able to restore through the GUI.
We have 5 instances on our clients old machine with SQL 2005, now they want us to move on cluster environment with SQL 2012.
Once I have installed and configured SQL cluster with single instance, but how to install SQL Cluster environment for Multiple instances.
Should I install all the instances first and than have to configure cluster ? OR Is there any way that first I will install cluster and than we can add the instances ?
What I have- Sql server 2012 (Standard Ed) Cluster on Windows 2012 R2 with both instances running on the same node- just to save on License, i.e. technically it’s Act/Pas cluster.
What I am looking for- how to configure cluster (e.g. via quorum, etc) to force both instances failed together? Means if for some reason 1-st instance will fail to node 2 another instance should follow (otherwise it will be Act/Act cluster and 2-nd license is required).
If there is no standard way (cluster configuration I mean) to do it I should create some custom process to monitor where each instance is running.
Curious what the industry standard is combining multiple instances on single server ?
Right now i have separate servers for OLTP, SSAS, and SSIS, and for OTP have Development, Test, and production environments. Im considering combining the SSIS and SSAS services into additional instances on each environment servers.
Now (Production) OLTP SSAS SSIS
After
1 Server with 3 separate instances for SSAS, SSIS, OLTP. Then replicate this model through the environments to utilize development and testing.
I am in the process of writing a generic PowerShell script (that executes T-SQL statements) to copy all or some databases from one SQL Server 2012 SP2 instance to a second instance that runs SQL Server 2012 SP2 or higher. The script is designed to be used on demand, and part of its execution will involve reconciling the logins between the source and destination instances.
I am looking for a reliable way to reconcile the logins between the instances in a somewhat repetitive (i.e., "on demand") fashion. There is no SSIS in the environment; so the SSIS transfer logins task to not available to me.
We have a SQL 2012 server instance that has log shipping set up to another SQL 2012 server to provide a warm standby for a forward facing application. The databases on the primary server occasionally are required to be backed up and restored to a development environment, completely different server. Is there a way to schedule full backups with log shipping enabled?
I am in plan to implement following for backup of one of our database Enable Full recovery mode
1- Create full backup nightly 2- Create transaction log backup after every 25 min
as I am taking full backup every night, I think I can remove transaction log file backups at the time of full backup, as we can apply transaction log backup over full backup.My question is regarding removal of transaction log backups.
-Should I remove all transaction log backups and then execute full backup? -Should I execute full backup and remove all transaction log backup older than 24Hrs ? -Do I have to consider SCN or related info before deleting any transaction log backup ?
I was told that if I wanted to setup mirroring between multiple instance on the same server that each endpoint had to be different number. Is that true? Or can I use the default 5022 for each instance?
Is there any Best Practice, 3rd Party Tool for Automating SQL Server Patching for around 10000+ SQL instances.
It is very difficult to do the same manually because each server will consume atleast 1-2 hours and if this is completely done manually then n number of people might be required for accomplishing this task which doesn't looks feasible at all.
It is an active passive cluster which doesn’t allow any testing. All instances have to be failed over together, we aren’t allowed to just failover 1 even for testing purposes. Node 1 is the active node and we can failover to node 2 for 30 days free of charge but services have to then be failed back.
We need to run the cluster with node 1 as the primary node always and 2 just use for failover testing or for less than 30 day periods whilst performing cluster patch upgrades etc.
Now l am sure we could fail over 1 instances at a time for testing and diagnosing issues plus if add a new instance that's not production to get to the platform level as the rest of the instances this would avoid taking production down in the fail over process.
I am currently having publisher(database A), subscriber (database B) and distributor on the same instance for a test environment that is using a uni directional transactional replication. Now I need to setup another unidirectional transactional replication in the same test instance but for a different database. Publisher database is D and subcsirber database is E.
