Jump to content


incorrect md5 for every checkpoint

md5

  • Please log in to reply
1 reply to this topic

#1 ownself

ownself

    Newbie

  • Members
  • Pip
  • 1 posts

Posted 05 May 2019 - 03:09 PM

Hi, guys

I really need some help here, anything you guys suggest I would appreciate.

We've had a local P4 server running for 4 years with 30 people team, but unfortunately we've not done even single time backup process, we did change our P4 from a small hard disk to a large disk 1 year and half ago, but with copy the whole P4 folder to new disk instead of official backup/restore process.

Right now our P4 folder is almost 800GB big, and db.have is 9GB, db.integed is 1.5GB, db.rev is 1.5GB. These are the 3 biggest db files. Recently some weird things happened including P4 service stopped from no reason(our P4 is running on Windows7), random file corrupted with one byte changed. And since our P4 server has run for such long time, we decide to change it to a new machine. This time we were thinking about using official backup process.

Well, then every time I did "p4d -jc" that getting a checkpoint file with different MD5, OK, maybe P4D input some date info into checkpoint file, but the checkpoint's md5 is not even same with the value in the md5 file(for saying, checkpoint.1 and checkpoint.1.md5)

I run HD Tune pro checking the hard disk without any problem, So, I have no clue what should I do for the next step.

Any suggestion? I'm begging you guys.

#2 Dave Foglesong

Dave Foglesong

    Member

  • Members
  • PipPip
  • 12 posts

Posted 09 May 2019 - 12:48 AM

> Well, then every time I did "p4d -jc" that getting a checkpoint file with different MD5, OK, maybe P4D input some date info into checkpoint file,

This is normal. Even if the server is shut down, when you create a checkpoint it updates some data (such as the "journal" and "lastCheckpointAction" counters) so each checkpoint will differ and have different digest values.

> but the checkpoint's md5 is not even same with the value in the md5 file(for saying, checkpoint.1 and checkpoint.1.md5)

This is not normal. The only time I've seen something like this was due to an oddity of expanding a compressed checkpoint on Windows. See the "troubleshooting" section at the bottom of this note for details: https://community.pe.../s/article/3034


If you're not using compressed checkpoints, then I'd look into something like a virus scanner, disk compression, or something else that might be changing the checkpoint file in some way. Or maybe try a different md5 digest calculation program. (Newer versions of Windows support the Linux subsystem, and you should be able to run the standard "md5sum" program there.)

Don't forget that the checkpoint file is only half the system. You also need to move all the archive files to the new server. Be sure to run "p4 verify" (see the admin guide) to make sure the archive files are all OK before migrating to a new server.





Also tagged with one or more of these keywords: md5

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users