Solved

MWS backup taking too long

  • 10 January 2022
  • 4 replies
  • 381 views

Userlevel 3
Badge +5

Hi all,

APPs 10 Upd 11.

It use to be the case that running the ifshome backup prior to installing a new delivery would take around 40 mins to complete by running the create backup script as admin.

However, now it is taking over an hour to complete for a 20gb backup. After it passes the hour mark, this message appears and no backup is created.

 

SEVERE: Panic message: Watchdog kicked in, operation takes too long time! Aborting…

 

So, increased the watchdog timeout limit to 2 hours and now the backup is created but it is incomplete as it hits the 2 hour mark and the same error message appears.

 

I dont think a mws backup should take this long to run seeing as the db backup completes in far less time that this.

So just wondering what would be considered a “reasonable time” and any suggestions to fix?

 

Thanks

 

 

 

 

icon

Best answer by Charith Epasinghe 10 January 2022, 10:41

View original

This topic has been closed for comments

4 replies

Userlevel 6
Badge +13

Hi @DBowden,

It is noted that you have already tried the inceasing the WATCHDOG_TIMEOUT.Watchdog error  is not a bug. rather an infrastructure related issue.

When the operation takes a long time, the Watch Dog Timer kicks in and the error message is thrown. 


Few suggestions :

1.  It is noted that you have altready done this.

You can set the watchdog timeout in mws.properties file (in ifshome/instance/<instance>/bin dir) by adding the line:
WATCHDOG_TIMEOUT=7200

2. Minimize the size of the backup

You can control what to include or exclude in the backup zip by using IncludeList and ExcludeList. Please refer to the "Including user-defined files in a backup or a clone" section in following documentation for more information:

https://wit.ifsworld.com/f1docs/apps10upd7/foundation1/020_installation/020_installing_fresh_system/040_installing_ifs_applications/030_deploy_middle_tier/hidden_itd/270_backup.htm.

Check whether there is a folder named "mws_bak" created inside the mws_home. If it exists, remove it . This backup folder is created during UPDs when changes are deployed to the Middleware Home.
 

3. Watchdog kicked error could be shown when the available resources in the VM is not enough for the installation process to complete. 

- If possible, Increase the available RAM, free disk space and number of virtual processors on the VM. 
-If multiple IFS environments are running on the same VM, shut down the other instances before proceed with tha backup.
 

Hope this answers your question.

Userlevel 6
Badge +18

@DBowden: Few things to look at,

  1. Check why is the MWS backup is so big as per Charith’s item # 2.
  2. You can use utilities like Treesize to review the instance home directory and figure out which sub folders are large and whether there are any log files which can be dropped
  3. If your infrastructure is virtualized, have you considered using VM backup/snapshot of the Application (MWS) Server instead of IFS Backup Utility as this is lot quicker to take a backup (and restore if needed).
Badge +1

I would also suggest to look to see if its also backing up ‘unwanted’ log files from the middleware… If so, remove / delete these log files prior to backing up.

Badge +3

Hi @Charith Epasinghe 
Tell me please,
Can I change the location of backup zip file instead of <ifs_home>/instance/<instance>?
I can't find this option in mws.properties