nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Time taken to re-index

Rahul Aich [Nagra]

Hi All

Do we have any guidelines available on how long it should take to re-index a jira instance?

My jira instance (4.1.2) takes 30 min to re-index (40,000 issues,2K users, 100 custom fields and 200 projects) so i am not quite sure if it is taking usual time or longer than expected?

Moreover, has anyone experienced faster index time after upgrading to a higher Jira version?

Rahul

Accepted answers

Nic Brough -Adaptavist-

Yes, I did a test recently - indexing an old v4 live system takes X minutes. Upgrading to 5.1, the reindexing took around 75% of the time.

However, there's no sane way to give you a quantitative guideline, because it's a function of

Your database access speed
How fast you can read/write the index disk
How much resource the server has (RAM/CPU etc)
Complexity of issue data - if you've no custom fields, it only needs to index the system fields. If you add custom field data, it needs to extract each piece of it and index that

And of course, there's an interaction where some of these will be utterly irrelevant to the speed because others are far slower and hence they're blocking/waiting.

However, for any given system, it does tend to scale roughly linearly with the number of issues. If you make no changes to your Jira, other than doubling the number of issues, you'll find it takes an hour to re-index. Delete 20,000 issues, you'll find it drops to quarter of an hour.

All comments

Nic Brough -Adaptavist-

Yes, I did a test recently - indexing an old v4 live system takes X minutes. Upgrading to 5.1, the reindexing took around 75% of the time.

However, there's no sane way to give you a quantitative guideline, because it's a function of

Your database access speed
How fast you can read/write the index disk
How much resource the server has (RAM/CPU etc)
Complexity of issue data - if you've no custom fields, it only needs to index the system fields. If you add custom field data, it needs to extract each piece of it and index that

And of course, there's an interaction where some of these will be utterly irrelevant to the speed because others are far slower and hence they're blocking/waiting.

Rahul Aich [Nagra]

Thanks Nic for your response. It was very informative.

Rahul

MatthewC

I think Nic has covered most of the points but asking a question like this it would also help to say what the type of environment was. There can be quite a bit of difference between a VM hosted service with data on a SAN vs a dedicated physical host with fast disks locally mounted.

But I'd say 30 min is fine for a server that size. The last one I reindexed was about 60k issue, 300 projects, 400 custom fields, 3k users, VM hosting with slow-ish disks and it took just over 30 mins

Rahul Aich [Nagra]

Thanks Matthew....

Athanasios Bladis

Hello,

i have the same issue, very slow reindx, my question ist, how to fix it? Jira is more than 6 hours needed to reindex.

619 Projects

7000 Users

16 Core Processor,

64 GB RAM

BR.

Sakis

BMW

Andre Lehmann

Our record:

260 projects
150000 issues
2000 users
220 workflows
220 custom-fields

VM: 4 CPUs, 4GB JVM
background-Re-Indexing: 9,5h
Foreground-Indexing: 75minutes

EXACT Same data on physical Server:
background-Re-Indexing: 4h
foreground-Indexing: 35minutes

Nic Brough -Adaptavist-

You don't tell us the volume of issues, so there's no way to tell if that's fast or slow.

Remember a reindex basically has to read every. single. issue. And all its data from several other tables. Then build and write an index record.

How to fix it depends on the volume and what the bottleneck might be, if there even is one.