- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
If someone can please comment,
We are running SAS 9.4 on Linux. Have below error in metadata server log, also, we don't see successful backup from sas management console,
:sas2 - Metadata Server Backup Not Executed Due To I/O Errors
what would be the most common cause for below error?
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hello @woo,
I have seen this before.
First, you can try to delete the journal backup file (then xml on your backup directory), so it will be recreated. Only CON, you won't be able to rollback that previous with the SMC anymore, since it will get a new registry of backups.
Also, perhaps you could give a look into this Note, although I doubt that will be your problem http://support.sas.com/kb/56/230.html
Finally, when you can turn SAS services down (evening or patch time), perhaps you can try to extend information in the logs on the metadata server, start it and run a manual backup and see if there is any additional information. When you get the logs, you should rollback the changes on the log configuration and restart tmetadata server again.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Do you have enough disk space where you are writing the backups to ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
yes, we checked that already, we do have enough space. specific directory is about 70% usage.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Has this ever worked ?
Can you maybe publish more of the log if you can ?
Are you running this backup manually from SMC ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
It is auto scheduled back up runs everyday at 1 am. it has stopped today with error mentioned earlier. some more errors are like,
Journal file calculatedSize=16 and lDataSize=0 and calculatedSize should never exceed lDataSize! Assuming calculation is correct and resetting lDataSize to match.
Journal file calculatedSize=1528 and lDataSize=1544 and the two should always be the same!
Crash recovery apparently failed.
An error occurred while attempting to Recover the Journal File after I/O failures.
Journal reload has failed.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
set as problematic ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
we can always do but just trying to find the root cause. Also, wanted to see if someone came across similar error and if can add any comment. Thanks...
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
No, unfortunately haven't seen this before.
My only suggestions would be to:
1.) Roll back the Metadata repository to at least get you back to a stable state
2.) Trace the metadata server logs line by line to see where the inconsistencies originated from
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Have you attempted a metadata Analyze&Repair from SAS Management Console? If not have a go and let us know what happened.
Regards,
- Jan.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hello @woo,
I have seen this before.
First, you can try to delete the journal backup file (then xml on your backup directory), so it will be recreated. Only CON, you won't be able to rollback that previous with the SMC anymore, since it will get a new registry of backups.
Also, perhaps you could give a look into this Note, although I doubt that will be your problem http://support.sas.com/kb/56/230.html
Finally, when you can turn SAS services down (evening or patch time), perhaps you can try to extend information in the logs on the metadata server, start it and run a manual backup and see if there is any additional information. When you get the logs, you should rollback the changes on the log configuration and restart tmetadata server again.