Page not reloading after quick posting

jamesd

The usual reload of a page to show the posted comment after a quick or other reply is posted isn't happening for me at the moment. Maybe a time-related glitch?

No problem to see the posted comment if I reload the thread in another window so the post itself does go through.

jamesd

And it also happens when creating a new topic with an initial post. In the window where I wrote the initial post it's still "Waiting for forums.moneysavingexpert.com" on the status bar even though I'm here replying to that post now.

jamesd

Same when editing a post and posting the changed post.

jamesd

But that last post did reload eventually, even though I've an edit one that hasn't reloaded for a good deal longer.

jamesd

Took just 15 seconds to reload after that last post.

Looking in some ways like a lagged slave check that's going wrong somewhere.

jamesd

Seems to be board-related in some way, over in Pensions it's either never reloading or exhausting how long I'll leave a window open before giving up - many minutes at least.

jamesd

While that post refreshed, over in Pensions I've still got one from earlier that hasn't.

jamesd

Just had one that reloaded over in pensions to the duplicate post warning. The post was shown in the topic review. Unlikely that I really did try duplicate posting.

Since I mentioned time, yes, my computer has adjusted to the correct winter time.

Meanwhile the Pensions post I mentioned in post #7 still hasn't reloaded.

jamesd

Now working normally for me including in the pension section.

torbrex

Well jamesd, it looks like you have been suffering this morning what some of us have been suffering for over 3 months, the dreaded server back-up slowdown.

I have started numerous threads about it but with no-one else complaining, I have been a lone voice in the wilderness.

Latest thread.

jamesd

There's not usually a lot of reason for that to be the cause since it's usually not too hard to avoid it.

One easy way is to use a MySQL replication slave. Stop the replication SQL thread and take the backup. While the SQL thread is stopped the IO thread will still be collecting updates from the master for use in disaster or other recovery. Once the backup is complete restart the replication SQL thread. Once that is caught up either put the slave back into normal slave service or if it's a spare for emergencies just let it continue to be that.

There can be technical or cost reasons not to do this but it tends to be reasonably straightforward. There are also a range of backup method choices that can be used and some of those are less likely to cause trouble than others. One way that can work quite well is an LVM snapshot, though only for InnoDB.

If for some reason that isn't practical, defragmenting can often help with the speed, even on Linux systems that have traditionally but not particularly correctly in this case often been regarded as not subject to fragmentation issues in the fileystem.