Long waiting time (and no success) for Send/Receive

Question

A colleague of mine is having a problem where going to File > Send/Receive Projects in Paratext 7.6(.54.139) results in a long waiting time (think of 10 minutes or more).
This is followed either by an error message ‘An internal error has occurred’ (HgCommandException, line 67 in lock), or by nothing (the ‘Loading data’ progress bar disappears).

Is it likely that there is a particular project or setting that is causing this?
And what would be the best way to get past this situation?

较早的帖子 - 以原始语言显示

Answer 1

Please see FB 58127.
This is a case right after the S/R “Loading data” progress bar had run for about a minute or so (without success).

The case where the “An internal error has occurred” error message was given, occurred on Jan 10 around 8:40 AM UTC.
(The message also said “Support has been notified”.)

较早的帖子 - 以原始语言显示

Comments

The delay is caused by the user settings project being locked - it looks
like the project files were copied from one virtual machine to another when
this project was locked.

There are two lock files that Mercurial uses:
{projectFolder}.hg\store\lock
{projectFolder}.hg\wlock

Usually both of these are deleted at the end of a command, but sometimes
they are left around and need to be manually deleted.

The user settings project was something we created in Paratext 7.6, but
decided to get rid of in Paratext 8. It’s the last project processed during
send/receive, so errors in it don’t prevent the other projects from being
completed.

The user settings project name is a hash value based on the user name so
usually looks like some random string - in the case of the reported
problem, it is l4Bwgh_BX4. Either of the two files mentioned above that
exist in this project will fix the problem the user is having.

John+Wickberg
Paratext Support

较早的帖子 - 以原始语言显示

---

There was indeed a wlock file, and deleting it resolved the issue. Thanks.
(The issue occurred on a somewhat special, server-based setup, so I guess you might be right about the file having been copied.)

较早的帖子 - 以原始语言显示

Answer 2

It is likely to be an issue with a project. Can you have him use Help > Report a problem right after the problem occurs so that we can try figure out the issue?

较早的帖子 - 以原始语言显示

Answer 3

I also have a user who can’t run Send/Receive right now. They haven’t indicated that there is an error message, but that it gets “stuck” partway through. (I haven’t seen the problem myself.) But I asked them about the lock files, and they said that both of these files are present:

{projectFolder}.hg\store\lock
{projectFolder}.hg\wlock

So would it be accurate to say that if you are not currently doing a Send/Receive, those two files should not be present, and can be safely deleted?

较早的帖子 - 以原始语言显示

Answer 4

I was hoping to get a quick reply on this…

If a user is not currently doing a Send/Receive, can those two lock files be safely deleted?

较早的帖子 - 以原始语言显示

Comments

I think you’re safe to delete those. On Jan 11th John+Wickberg said:

<<There are two lock files that Mercurial uses:
{projectFolder}.hg\store\lock
{projectFolder}.hg\wlock

Usually both of these are deleted at the end of a command, but sometimes
they are left around and need to be manually deleted.>>

较早的帖子 - 以原始语言显示

---

jeffh:

I was hoping to get a quick reply on this…

If a user is not currently doing a Send/Receive, can those two lock files be safely deleted?

Sorry for the slow response.
I would take it a step further, though, to be safe: If Paratext is not , it should be safe to delete those files. There are other operations Paratext does besides S/R that can lock the Mercurial repository.

较早的帖子 - 以原始语言显示